Index: llvm/trunk/test/CodeGen/PowerPC/vector-constrained-fp-intrinsics.ll =================================================================== --- llvm/trunk/test/CodeGen/PowerPC/vector-constrained-fp-intrinsics.ll +++ llvm/trunk/test/CodeGen/PowerPC/vector-constrained-fp-intrinsics.ll @@ -0,0 +1,10810 @@ +; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py +; RUN: llc -O3 -mtriple=powerpc64le-linux-gnu < %s | FileCheck --check-prefix=PC64LE %s +; RUN: llc -O3 -mtriple=powerpc64le-linux-gnu -mcpu=pwr9 < %s | FileCheck --check-prefix=PC64LE9 %s +; RUN: llc -O3 -mtriple=powerpc64-linux-gnu < %s | FileCheck --check-prefix=PC64 %s + +define <1 x float> @constrained_vector_fdiv_v1f32() { +; PC64LE-LABEL: constrained_vector_fdiv_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI0_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI0_1@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI0_0@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI0_1@toc@l(4) +; PC64LE-NEXT: xsdivsp 0, 1, 0 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fdiv_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI0_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI0_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI0_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI0_1@toc@l(3) +; PC64LE9-NEXT: xsdivsp 0, 1, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fdiv_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI0_0@toc@ha +; PC64-NEXT: lfs 0, .LCPI0_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI0_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI0_1@toc@l(3) +; PC64-NEXT: fdivs 1, 1, 0 +; PC64-NEXT: blr +entry: + %div = call <1 x float> @llvm.experimental.constrained.fdiv.v1f32( + <1 x float> , + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %div +} + +define <2 x double> @constrained_vector_fdiv_v2f64() { +; PC64LE-LABEL: constrained_vector_fdiv_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI1_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI1_1@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI1_0@toc@l +; PC64LE-NEXT: addi 4, 4, .LCPI1_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: lxvd2x 1, 0, 4 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xvdivdp 34, 1, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fdiv_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI1_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI1_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI1_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI1_1@toc@l +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: xvdivdp 34, 1, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fdiv_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI1_0@toc@ha +; PC64-NEXT: lfs 0, .LCPI1_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI1_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI1_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI1_2@toc@ha +; PC64-NEXT: lfs 2, .LCPI1_2@toc@l(3) +; PC64-NEXT: fdiv 1, 1, 0 +; PC64-NEXT: fdiv 2, 2, 0 +; PC64-NEXT: blr +entry: + %div = call <2 x double> @llvm.experimental.constrained.fdiv.v2f64( + <2 x double> , + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %div +} + +define <3 x float> @constrained_vector_fdiv_v3f32() { +; PC64LE-LABEL: constrained_vector_fdiv_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI2_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI2_3@toc@ha +; PC64LE-NEXT: addis 5, 2, .LCPI2_2@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI2_0@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI2_3@toc@l(4) +; PC64LE-NEXT: lfs 2, .LCPI2_2@toc@l(5) +; PC64LE-NEXT: addis 3, 2, .LCPI2_1@toc@ha +; PC64LE-NEXT: xsdivsp 1, 1, 0 +; PC64LE-NEXT: lfs 3, .LCPI2_1@toc@l(3) +; PC64LE-NEXT: addis 3, 2, .LCPI2_4@toc@ha +; PC64LE-NEXT: xsdivsp 2, 2, 0 +; PC64LE-NEXT: addi 3, 3, .LCPI2_4@toc@l +; PC64LE-NEXT: lvx 4, 0, 3 +; PC64LE-NEXT: xsdivsp 0, 3, 0 +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: xscvdpspn 2, 2 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 1, 1, 1 +; PC64LE-NEXT: xxsldwi 35, 2, 2, 1 +; PC64LE-NEXT: vmrglw 2, 3, 2 +; PC64LE-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 3, 2, 4 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fdiv_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI2_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI2_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI2_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI2_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI2_2@toc@ha +; PC64LE9-NEXT: xsdivsp 1, 1, 0 +; PC64LE9-NEXT: lfs 2, .LCPI2_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI2_3@toc@ha +; PC64LE9-NEXT: lfs 3, .LCPI2_3@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI2_4@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI2_4@toc@l +; PC64LE9-NEXT: lxvx 36, 0, 3 +; PC64LE9-NEXT: xsdivsp 2, 2, 0 +; PC64LE9-NEXT: xsdivsp 0, 3, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 2 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 3, 2, 4 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fdiv_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI2_0@toc@ha +; PC64-NEXT: lfs 0, .LCPI2_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI2_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI2_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI2_2@toc@ha +; PC64-NEXT: lfs 2, .LCPI2_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI2_3@toc@ha +; PC64-NEXT: lfs 3, .LCPI2_3@toc@l(3) +; PC64-NEXT: fdivs 1, 1, 0 +; PC64-NEXT: fdivs 2, 2, 0 +; PC64-NEXT: fdivs 3, 3, 0 +; PC64-NEXT: blr +entry: + %div = call <3 x float> @llvm.experimental.constrained.fdiv.v3f32( + <3 x float> , + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %div +} + +define <3 x double> @constrained_vector_fdiv_v3f64() { +; PC64LE-LABEL: constrained_vector_fdiv_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI3_2@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI3_3@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI3_2@toc@l +; PC64LE-NEXT: addi 4, 4, .LCPI3_3@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: lxvd2x 1, 0, 4 +; PC64LE-NEXT: addis 3, 2, .LCPI3_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI3_1@toc@ha +; PC64LE-NEXT: lfs 3, .LCPI3_0@toc@l(3) +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xvdivdp 2, 1, 0 +; PC64LE-NEXT: lfs 0, .LCPI3_1@toc@l(4) +; PC64LE-NEXT: xsdivdp 3, 0, 3 +; PC64LE-NEXT: xxswapd 1, 2 +; PC64LE-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fdiv_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI3_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI3_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI3_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI3_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI3_2@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI3_2@toc@l +; PC64LE9-NEXT: xsdivdp 3, 1, 0 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI3_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI3_3@toc@l +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: xvdivdp 2, 1, 0 +; PC64LE9-NEXT: xxswapd 1, 2 +; PC64LE9-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE9-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fdiv_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI3_0@toc@ha +; PC64-NEXT: lfs 0, .LCPI3_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI3_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI3_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI3_2@toc@ha +; PC64-NEXT: lfs 2, .LCPI3_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI3_3@toc@ha +; PC64-NEXT: lfs 3, .LCPI3_3@toc@l(3) +; PC64-NEXT: fdiv 1, 1, 0 +; PC64-NEXT: fdiv 2, 2, 0 +; PC64-NEXT: fdiv 3, 3, 0 +; PC64-NEXT: blr +entry: + %div = call <3 x double> @llvm.experimental.constrained.fdiv.v3f64( + <3 x double> , + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %div +} + +define <4 x double> @constrained_vector_fdiv_v4f64() { +; PC64LE-LABEL: constrained_vector_fdiv_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI4_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI4_1@toc@ha +; PC64LE-NEXT: addis 5, 2, .LCPI4_2@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI4_0@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: addi 3, 4, .LCPI4_1@toc@l +; PC64LE-NEXT: addi 4, 5, .LCPI4_2@toc@l +; PC64LE-NEXT: lxvd2x 1, 0, 3 +; PC64LE-NEXT: lxvd2x 2, 0, 4 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xxswapd 2, 2 +; PC64LE-NEXT: xvdivdp 34, 1, 0 +; PC64LE-NEXT: xvdivdp 35, 2, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fdiv_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI4_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI4_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI4_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI4_1@toc@l +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI4_2@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI4_2@toc@l +; PC64LE9-NEXT: xvdivdp 34, 1, 0 +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: xvdivdp 35, 1, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fdiv_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI4_0@toc@ha +; PC64-NEXT: lfs 0, .LCPI4_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI4_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI4_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI4_2@toc@ha +; PC64-NEXT: lfs 2, .LCPI4_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI4_3@toc@ha +; PC64-NEXT: lfs 3, .LCPI4_3@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI4_4@toc@ha +; PC64-NEXT: fdiv 1, 1, 0 +; PC64-NEXT: lfs 4, .LCPI4_4@toc@l(3) +; PC64-NEXT: fdiv 2, 2, 0 +; PC64-NEXT: fdiv 3, 3, 0 +; PC64-NEXT: fdiv 4, 4, 0 +; PC64-NEXT: blr +entry: + %div = call <4 x double> @llvm.experimental.constrained.fdiv.v4f64( + <4 x double> , + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %div +} + +define <1 x float> @constrained_vector_frem_v1f32() { +; PC64LE-LABEL: constrained_vector_frem_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI5_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI5_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI5_0@toc@l(3) +; PC64LE-NEXT: lfs 2, .LCPI5_1@toc@l(4) +; PC64LE-NEXT: bl fmodf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_frem_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI5_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI5_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI5_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI5_1@toc@l(3) +; PC64LE9-NEXT: bl fmodf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_frem_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI5_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI5_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI5_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI5_1@toc@l(3) +; PC64-NEXT: bl fmodf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %rem = call <1 x float> @llvm.experimental.constrained.frem.v1f32( + <1 x float> , + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %rem +} + +define <2 x double> @constrained_vector_frem_v2f64() { +; PC64LE-LABEL: constrained_vector_frem_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: addis 4, 2, .LCPI6_1@toc@ha +; PC64LE-NEXT: addis 3, 2, .LCPI6_0@toc@ha +; PC64LE-NEXT: lfs 31, .LCPI6_1@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI6_0@toc@l(3) +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: bl fmod +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI6_2@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI6_2@toc@l(3) +; PC64LE-NEXT: bl fmod +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 1, 0 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_frem_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI6_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI6_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI6_1@toc@ha +; PC64LE9-NEXT: lfs 31, .LCPI6_1@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl fmod +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI6_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI6_2@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl fmod +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_frem_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI6_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI6_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI6_1@toc@ha +; PC64-NEXT: lfs 31, .LCPI6_1@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmod +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI6_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI6_2@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmod +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %rem = call <2 x double> @llvm.experimental.constrained.frem.v2f64( + <2 x double> , + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %rem +} + +define <3 x float> @constrained_vector_frem_v3f32() { +; PC64LE-LABEL: constrained_vector_frem_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f29, -24 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 29, -24(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: addis 4, 2, .LCPI7_1@toc@ha +; PC64LE-NEXT: addis 3, 2, .LCPI7_0@toc@ha +; PC64LE-NEXT: lfs 31, .LCPI7_1@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI7_0@toc@l(3) +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: bl fmodf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI7_2@toc@ha +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI7_2@toc@l(3) +; PC64LE-NEXT: bl fmodf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI7_3@toc@ha +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: fmr 29, 1 +; PC64LE-NEXT: lfs 1, .LCPI7_3@toc@l(3) +; PC64LE-NEXT: bl fmodf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 29 +; PC64LE-NEXT: addis 3, 2, .LCPI7_4@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI7_4@toc@l +; PC64LE-NEXT: lvx 4, 0, 3 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 3, 2, 4 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 29, -24(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_frem_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f29, -24 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 29, -24(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI7_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI7_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI7_1@toc@ha +; PC64LE9-NEXT: lfs 31, .LCPI7_1@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl fmodf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI7_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI7_2@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl fmodf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI7_3@toc@ha +; PC64LE9-NEXT: fmr 29, 1 +; PC64LE9-NEXT: lfs 1, .LCPI7_3@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl fmodf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 29 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: addis 3, 2, .LCPI7_4@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI7_4@toc@l +; PC64LE9-NEXT: lxvx 36, 0, 3 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 3, 2, 4 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 29, -24(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_frem_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI7_0@toc@ha +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI7_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI7_1@toc@ha +; PC64-NEXT: lfs 31, .LCPI7_1@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmodf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI7_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI7_2@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmodf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI7_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfs 1, .LCPI7_3@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmodf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 30 +; PC64-NEXT: fmr 2, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %rem = call <3 x float> @llvm.experimental.constrained.frem.v3f32( + <3 x float> , + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %rem +} + +define <3 x double> @constrained_vector_frem_v3f64() { +; PC64LE-LABEL: constrained_vector_frem_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -96(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 96 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: .cfi_offset v31, -32 +; PC64LE-NEXT: addis 4, 2, .LCPI8_1@toc@ha +; PC64LE-NEXT: stfd 31, 88(1) # 8-byte Folded Spill +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI8_0@toc@ha +; PC64LE-NEXT: lfs 31, .LCPI8_1@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI8_0@toc@l(3) +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: bl fmod +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI8_2@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI8_2@toc@l(3) +; PC64LE-NEXT: bl fmod +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI8_3@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfs 1, .LCPI8_3@toc@l(3) +; PC64LE-NEXT: bl fmod +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: lfd 31, 88(1) # 8-byte Folded Reload +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 96 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_frem_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -80(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 80 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: .cfi_offset v31, -32 +; PC64LE9-NEXT: addis 3, 2, .LCPI8_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI8_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI8_1@toc@ha +; PC64LE9-NEXT: stfd 31, 72(1) # 8-byte Folded Spill +; PC64LE9-NEXT: lfs 31, .LCPI8_1@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl fmod +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI8_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI8_2@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl fmod +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI8_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfs 1, .LCPI8_3@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl fmod +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: lfd 31, 72(1) # 8-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 80 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_frem_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI8_0@toc@ha +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI8_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI8_1@toc@ha +; PC64-NEXT: lfs 31, .LCPI8_1@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmod +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI8_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI8_2@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmod +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI8_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfs 1, .LCPI8_3@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmod +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 30 +; PC64-NEXT: fmr 2, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %rem = call <3 x double> @llvm.experimental.constrained.frem.v3f64( + <3 x double> , + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %rem +} + +define <4 x double> @constrained_vector_frem_v4f64() { +; PC64LE-LABEL: constrained_vector_frem_v4f64: +; PC64LE: # %bb.0: +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -96(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 96 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: .cfi_offset v31, -32 +; PC64LE-NEXT: addis 4, 2, .LCPI9_1@toc@ha +; PC64LE-NEXT: stfd 31, 88(1) # 8-byte Folded Spill +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI9_0@toc@ha +; PC64LE-NEXT: lfs 31, .LCPI9_1@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI9_0@toc@l(3) +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: bl fmod +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI9_2@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI9_2@toc@l(3) +; PC64LE-NEXT: bl fmod +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI9_3@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfs 1, .LCPI9_3@toc@l(3) +; PC64LE-NEXT: bl fmod +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI9_4@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI9_4@toc@l(3) +; PC64LE-NEXT: bl fmod +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 2, 31 +; PC64LE-NEXT: lfd 31, 88(1) # 8-byte Folded Reload +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 35, 1, 0 +; PC64LE-NEXT: addi 1, 1, 96 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_frem_v4f64: +; PC64LE9: # %bb.0: +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -80(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 80 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: .cfi_offset v31, -32 +; PC64LE9-NEXT: addis 3, 2, .LCPI9_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI9_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI9_1@toc@ha +; PC64LE9-NEXT: stfd 31, 72(1) # 8-byte Folded Spill +; PC64LE9-NEXT: lfs 31, .LCPI9_1@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl fmod +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI9_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI9_2@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl fmod +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI9_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfs 1, .LCPI9_3@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl fmod +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI9_4@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI9_4@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl fmod +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 2, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: lfd 31, 72(1) # 8-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 35, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 80 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_frem_v4f64: +; PC64: # %bb.0: +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f28, -32 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI9_0@toc@ha +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 28, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI9_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI9_1@toc@ha +; PC64-NEXT: lfs 31, .LCPI9_1@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmod +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI9_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI9_2@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmod +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI9_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfs 1, .LCPI9_3@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmod +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI9_4@toc@ha +; PC64-NEXT: fmr 28, 1 +; PC64-NEXT: lfs 1, .LCPI9_4@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmod +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 30 +; PC64-NEXT: fmr 2, 29 +; PC64-NEXT: fmr 3, 28 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 28, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr + %rem = call <4 x double> @llvm.experimental.constrained.frem.v4f64( + <4 x double> , + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %rem +} + +define <1 x float> @constrained_vector_fmul_v1f32() { +; PC64LE-LABEL: constrained_vector_fmul_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI10_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI10_1@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI10_0@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI10_1@toc@l(4) +; PC64LE-NEXT: xsmulsp 0, 1, 0 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fmul_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI10_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI10_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI10_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI10_1@toc@l(3) +; PC64LE9-NEXT: xsmulsp 0, 1, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fmul_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI10_0@toc@ha +; PC64-NEXT: lfs 0, .LCPI10_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI10_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI10_1@toc@l(3) +; PC64-NEXT: fmuls 1, 1, 0 +; PC64-NEXT: blr +entry: + %mul = call <1 x float> @llvm.experimental.constrained.fmul.v1f32( + <1 x float> , + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %mul +} + +define <2 x double> @constrained_vector_fmul_v2f64() { +; PC64LE-LABEL: constrained_vector_fmul_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI11_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI11_1@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI11_0@toc@l +; PC64LE-NEXT: addi 4, 4, .LCPI11_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: lxvd2x 1, 0, 4 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xvmuldp 34, 1, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fmul_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI11_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI11_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI11_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI11_1@toc@l +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: xvmuldp 34, 1, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fmul_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI11_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI11_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI11_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI11_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI11_2@toc@ha +; PC64-NEXT: lfs 2, .LCPI11_2@toc@l(3) +; PC64-NEXT: fmul 1, 0, 1 +; PC64-NEXT: fmul 2, 0, 2 +; PC64-NEXT: blr +entry: + %mul = call <2 x double> @llvm.experimental.constrained.fmul.v2f64( + <2 x double> , + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %mul +} + +define <3 x float> @constrained_vector_fmul_v3f32() { +; PC64LE-LABEL: constrained_vector_fmul_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI12_1@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI12_3@toc@ha +; PC64LE-NEXT: addis 5, 2, .LCPI12_2@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI12_1@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI12_3@toc@l(4) +; PC64LE-NEXT: lfs 2, .LCPI12_2@toc@l(5) +; PC64LE-NEXT: addis 3, 2, .LCPI12_0@toc@ha +; PC64LE-NEXT: xsmulsp 1, 0, 1 +; PC64LE-NEXT: lfs 3, .LCPI12_0@toc@l(3) +; PC64LE-NEXT: addis 3, 2, .LCPI12_4@toc@ha +; PC64LE-NEXT: xsmulsp 2, 0, 2 +; PC64LE-NEXT: addi 3, 3, .LCPI12_4@toc@l +; PC64LE-NEXT: lvx 4, 0, 3 +; PC64LE-NEXT: xsmulsp 0, 0, 3 +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: xscvdpspn 2, 2 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 1, 1, 1 +; PC64LE-NEXT: xxsldwi 35, 2, 2, 1 +; PC64LE-NEXT: vmrglw 2, 3, 2 +; PC64LE-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 3, 2, 4 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fmul_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI12_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI12_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI12_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI12_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI12_2@toc@ha +; PC64LE9-NEXT: xsmulsp 0, 1, 0 +; PC64LE9-NEXT: lfs 2, .LCPI12_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI12_3@toc@ha +; PC64LE9-NEXT: lfs 3, .LCPI12_3@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI12_4@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI12_4@toc@l +; PC64LE9-NEXT: lxvx 36, 0, 3 +; PC64LE9-NEXT: xsmulsp 2, 1, 2 +; PC64LE9-NEXT: xsmulsp 1, 1, 3 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xscvdpspn 1, 1 +; PC64LE9-NEXT: xxsldwi 34, 1, 1, 1 +; PC64LE9-NEXT: xscvdpspn 1, 2 +; PC64LE9-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 3, 2, 4 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fmul_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI12_0@toc@ha +; PC64-NEXT: lfs 0, .LCPI12_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI12_1@toc@ha +; PC64-NEXT: lfs 3, .LCPI12_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI12_2@toc@ha +; PC64-NEXT: lfs 2, .LCPI12_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI12_3@toc@ha +; PC64-NEXT: lfs 4, .LCPI12_3@toc@l(3) +; PC64-NEXT: fmuls 1, 3, 0 +; PC64-NEXT: fmuls 2, 3, 2 +; PC64-NEXT: fmuls 3, 3, 4 +; PC64-NEXT: blr +entry: + %mul = call <3 x float> @llvm.experimental.constrained.fmul.v3f32( + <3 x float> , + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %mul +} + +define <3 x double> @constrained_vector_fmul_v3f64() { +; PC64LE-LABEL: constrained_vector_fmul_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI13_2@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI13_3@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI13_2@toc@l +; PC64LE-NEXT: addi 4, 4, .LCPI13_3@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: lxvd2x 1, 0, 4 +; PC64LE-NEXT: addis 3, 2, .LCPI13_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI13_1@toc@ha +; PC64LE-NEXT: lfd 3, .LCPI13_0@toc@l(3) +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xvmuldp 2, 1, 0 +; PC64LE-NEXT: lfs 0, .LCPI13_1@toc@l(4) +; PC64LE-NEXT: xsmuldp 3, 3, 0 +; PC64LE-NEXT: xxswapd 1, 2 +; PC64LE-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fmul_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI13_0@toc@ha +; PC64LE9-NEXT: lfd 0, .LCPI13_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI13_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI13_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI13_2@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI13_2@toc@l +; PC64LE9-NEXT: xsmuldp 3, 0, 1 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI13_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI13_3@toc@l +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: xvmuldp 2, 1, 0 +; PC64LE9-NEXT: xxswapd 1, 2 +; PC64LE9-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE9-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fmul_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI13_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI13_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI13_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI13_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI13_2@toc@ha +; PC64-NEXT: lfs 2, .LCPI13_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI13_3@toc@ha +; PC64-NEXT: lfs 3, .LCPI13_3@toc@l(3) +; PC64-NEXT: fmul 1, 0, 1 +; PC64-NEXT: fmul 2, 0, 2 +; PC64-NEXT: fmul 3, 0, 3 +; PC64-NEXT: blr +entry: + %mul = call <3 x double> @llvm.experimental.constrained.fmul.v3f64( + <3 x double> , + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %mul +} + +define <4 x double> @constrained_vector_fmul_v4f64() { +; PC64LE-LABEL: constrained_vector_fmul_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI14_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI14_1@toc@ha +; PC64LE-NEXT: addis 5, 2, .LCPI14_2@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI14_0@toc@l +; PC64LE-NEXT: addi 4, 4, .LCPI14_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: addi 3, 5, .LCPI14_2@toc@l +; PC64LE-NEXT: lxvd2x 1, 0, 4 +; PC64LE-NEXT: lxvd2x 2, 0, 3 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xxswapd 2, 2 +; PC64LE-NEXT: xvmuldp 34, 1, 0 +; PC64LE-NEXT: xvmuldp 35, 1, 2 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fmul_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI14_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI14_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI14_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI14_1@toc@l +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI14_2@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI14_2@toc@l +; PC64LE9-NEXT: xvmuldp 34, 1, 0 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvmuldp 35, 1, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fmul_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI14_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI14_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI14_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI14_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI14_2@toc@ha +; PC64-NEXT: lfs 2, .LCPI14_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI14_3@toc@ha +; PC64-NEXT: lfs 3, .LCPI14_3@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI14_4@toc@ha +; PC64-NEXT: fmul 1, 0, 1 +; PC64-NEXT: lfs 4, .LCPI14_4@toc@l(3) +; PC64-NEXT: fmul 2, 0, 2 +; PC64-NEXT: fmul 3, 0, 3 +; PC64-NEXT: fmul 4, 0, 4 +; PC64-NEXT: blr +entry: + %mul = call <4 x double> @llvm.experimental.constrained.fmul.v4f64( + <4 x double> , + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %mul +} + +define <1 x float> @constrained_vector_fadd_v1f32() { +; PC64LE-LABEL: constrained_vector_fadd_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI15_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI15_1@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI15_0@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI15_1@toc@l(4) +; PC64LE-NEXT: xsaddsp 0, 1, 0 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fadd_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI15_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI15_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI15_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI15_1@toc@l(3) +; PC64LE9-NEXT: xsaddsp 0, 1, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fadd_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI15_0@toc@ha +; PC64-NEXT: lfs 0, .LCPI15_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI15_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI15_1@toc@l(3) +; PC64-NEXT: fadds 1, 1, 0 +; PC64-NEXT: blr +entry: + %add = call <1 x float> @llvm.experimental.constrained.fadd.v1f32( + <1 x float> , + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %add +} + +define <2 x double> @constrained_vector_fadd_v2f64() { +; PC64LE-LABEL: constrained_vector_fadd_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI16_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI16_1@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI16_0@toc@l +; PC64LE-NEXT: addi 4, 4, .LCPI16_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: lxvd2x 1, 0, 4 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xvadddp 34, 1, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fadd_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI16_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI16_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI16_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI16_1@toc@l +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: xvadddp 34, 1, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fadd_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI16_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI16_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI16_1@toc@ha +; PC64-NEXT: lfd 1, .LCPI16_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI16_2@toc@ha +; PC64-NEXT: lfs 3, .LCPI16_2@toc@l(3) +; PC64-NEXT: fadd 2, 1, 0 +; PC64-NEXT: fadd 1, 1, 3 +; PC64-NEXT: blr +entry: + %add = call <2 x double> @llvm.experimental.constrained.fadd.v2f64( + <2 x double> , + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %add +} + +define <3 x float> @constrained_vector_fadd_v3f32() { +; PC64LE-LABEL: constrained_vector_fadd_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI17_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI17_2@toc@ha +; PC64LE-NEXT: addis 5, 2, .LCPI17_1@toc@ha +; PC64LE-NEXT: xxlxor 3, 3, 3 +; PC64LE-NEXT: lfs 0, .LCPI17_0@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI17_2@toc@l(4) +; PC64LE-NEXT: lfs 2, .LCPI17_1@toc@l(5) +; PC64LE-NEXT: addis 3, 2, .LCPI17_3@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI17_3@toc@l +; PC64LE-NEXT: xsaddsp 1, 0, 1 +; PC64LE-NEXT: lvx 4, 0, 3 +; PC64LE-NEXT: xsaddsp 2, 0, 2 +; PC64LE-NEXT: xsaddsp 0, 0, 3 +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: xscvdpspn 2, 2 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 1, 1, 1 +; PC64LE-NEXT: xxsldwi 35, 2, 2, 1 +; PC64LE-NEXT: vmrglw 2, 3, 2 +; PC64LE-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 3, 2, 4 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fadd_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI17_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI17_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI17_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI17_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI17_2@toc@ha +; PC64LE9-NEXT: xsaddsp 2, 0, 2 +; PC64LE9-NEXT: lfs 3, .LCPI17_2@toc@l(3) +; PC64LE9-NEXT: xxlxor 1, 1, 1 +; PC64LE9-NEXT: xsaddsp 1, 0, 1 +; PC64LE9-NEXT: xsaddsp 0, 0, 3 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 2 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: addis 3, 2, .LCPI17_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI17_3@toc@l +; PC64LE9-NEXT: lxvx 36, 0, 3 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 3, 2, 4 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fadd_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI17_0@toc@ha +; PC64-NEXT: lfs 0, .LCPI17_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI17_1@toc@ha +; PC64-NEXT: lfs 3, .LCPI17_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI17_2@toc@ha +; PC64-NEXT: lfs 2, .LCPI17_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI17_3@toc@ha +; PC64-NEXT: lfs 4, .LCPI17_3@toc@l(3) +; PC64-NEXT: fadds 1, 3, 0 +; PC64-NEXT: fadds 2, 3, 2 +; PC64-NEXT: fadds 3, 3, 4 +; PC64-NEXT: blr +entry: + %add = call <3 x float> @llvm.experimental.constrained.fadd.v3f32( + <3 x float> , + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %add +} + +define <3 x double> @constrained_vector_fadd_v3f64() { +; PC64LE-LABEL: constrained_vector_fadd_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI18_1@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI18_2@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI18_1@toc@l +; PC64LE-NEXT: addi 4, 4, .LCPI18_2@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: lxvd2x 1, 0, 4 +; PC64LE-NEXT: addis 3, 2, .LCPI18_0@toc@ha +; PC64LE-NEXT: lfd 3, .LCPI18_0@toc@l(3) +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xvadddp 2, 1, 0 +; PC64LE-NEXT: xxlxor 0, 0, 0 +; PC64LE-NEXT: xsadddp 3, 3, 0 +; PC64LE-NEXT: xxswapd 1, 2 +; PC64LE-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fadd_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI18_0@toc@ha +; PC64LE9-NEXT: lfd 0, .LCPI18_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI18_1@toc@ha +; PC64LE9-NEXT: xxlxor 1, 1, 1 +; PC64LE9-NEXT: addi 3, 3, .LCPI18_1@toc@l +; PC64LE9-NEXT: xsadddp 3, 0, 1 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI18_2@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI18_2@toc@l +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: xvadddp 2, 1, 0 +; PC64LE9-NEXT: xxswapd 1, 2 +; PC64LE9-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE9-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fadd_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI18_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI18_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI18_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI18_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI18_2@toc@ha +; PC64-NEXT: lfs 2, .LCPI18_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI18_3@toc@ha +; PC64-NEXT: lfs 3, .LCPI18_3@toc@l(3) +; PC64-NEXT: fadd 1, 0, 1 +; PC64-NEXT: fadd 2, 0, 2 +; PC64-NEXT: fadd 3, 0, 3 +; PC64-NEXT: blr +entry: + %add = call <3 x double> @llvm.experimental.constrained.fadd.v3f64( + <3 x double> , + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %add +} + +define <4 x double> @constrained_vector_fadd_v4f64() { +; PC64LE-LABEL: constrained_vector_fadd_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI19_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI19_1@toc@ha +; PC64LE-NEXT: addis 5, 2, .LCPI19_2@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI19_0@toc@l +; PC64LE-NEXT: addi 4, 4, .LCPI19_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: addi 3, 5, .LCPI19_2@toc@l +; PC64LE-NEXT: lxvd2x 1, 0, 4 +; PC64LE-NEXT: lxvd2x 2, 0, 3 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xxswapd 2, 2 +; PC64LE-NEXT: xvadddp 34, 1, 0 +; PC64LE-NEXT: xvadddp 35, 1, 2 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fadd_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI19_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI19_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI19_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI19_1@toc@l +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI19_2@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI19_2@toc@l +; PC64LE9-NEXT: xvadddp 34, 1, 0 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvadddp 35, 1, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fadd_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI19_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI19_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI19_1@toc@ha +; PC64-NEXT: lfd 3, .LCPI19_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI19_2@toc@ha +; PC64-NEXT: lfd 1, .LCPI19_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI19_3@toc@ha +; PC64-NEXT: lfs 5, .LCPI19_3@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI19_4@toc@ha +; PC64-NEXT: fadd 2, 3, 0 +; PC64-NEXT: lfs 6, .LCPI19_4@toc@l(3) +; PC64-NEXT: fadd 4, 3, 1 +; PC64-NEXT: fadd 1, 3, 5 +; PC64-NEXT: fadd 3, 3, 6 +; PC64-NEXT: blr +entry: + %add = call <4 x double> @llvm.experimental.constrained.fadd.v4f64( + <4 x double> , + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %add +} + +define <1 x float> @constrained_vector_fsub_v1f32() { +; PC64LE-LABEL: constrained_vector_fsub_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI20_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI20_1@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI20_0@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI20_1@toc@l(4) +; PC64LE-NEXT: xssubsp 0, 1, 0 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fsub_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI20_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI20_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI20_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI20_1@toc@l(3) +; PC64LE9-NEXT: xssubsp 0, 1, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fsub_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI20_0@toc@ha +; PC64-NEXT: lfs 0, .LCPI20_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI20_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI20_1@toc@l(3) +; PC64-NEXT: fsubs 1, 1, 0 +; PC64-NEXT: blr +entry: + %sub = call <1 x float> @llvm.experimental.constrained.fsub.v1f32( + <1 x float> , + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %sub +} + +define <2 x double> @constrained_vector_fsub_v2f64() { +; PC64LE-LABEL: constrained_vector_fsub_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI21_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI21_1@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI21_0@toc@l +; PC64LE-NEXT: addi 4, 4, .LCPI21_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: lxvd2x 1, 0, 4 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xvsubdp 34, 1, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fsub_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI21_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI21_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI21_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI21_1@toc@l +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: xvsubdp 34, 1, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fsub_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI21_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI21_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI21_1@toc@ha +; PC64-NEXT: lfd 1, .LCPI21_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI21_2@toc@ha +; PC64-NEXT: lfs 3, .LCPI21_2@toc@l(3) +; PC64-NEXT: fsub 2, 1, 0 +; PC64-NEXT: fsub 1, 1, 3 +; PC64-NEXT: blr +entry: + %sub = call <2 x double> @llvm.experimental.constrained.fsub.v2f64( + <2 x double> , + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %sub +} + +define <3 x float> @constrained_vector_fsub_v3f32() { +; PC64LE-LABEL: constrained_vector_fsub_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI22_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI22_2@toc@ha +; PC64LE-NEXT: addis 5, 2, .LCPI22_1@toc@ha +; PC64LE-NEXT: xxlxor 3, 3, 3 +; PC64LE-NEXT: lfs 0, .LCPI22_0@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI22_2@toc@l(4) +; PC64LE-NEXT: lfs 2, .LCPI22_1@toc@l(5) +; PC64LE-NEXT: addis 3, 2, .LCPI22_3@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI22_3@toc@l +; PC64LE-NEXT: xssubsp 1, 0, 1 +; PC64LE-NEXT: lvx 4, 0, 3 +; PC64LE-NEXT: xssubsp 2, 0, 2 +; PC64LE-NEXT: xssubsp 0, 0, 3 +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: xscvdpspn 2, 2 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 1, 1, 1 +; PC64LE-NEXT: xxsldwi 35, 2, 2, 1 +; PC64LE-NEXT: vmrglw 2, 3, 2 +; PC64LE-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 3, 2, 4 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fsub_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI22_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI22_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI22_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI22_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI22_2@toc@ha +; PC64LE9-NEXT: xssubsp 2, 0, 2 +; PC64LE9-NEXT: lfs 3, .LCPI22_2@toc@l(3) +; PC64LE9-NEXT: xxlxor 1, 1, 1 +; PC64LE9-NEXT: xssubsp 1, 0, 1 +; PC64LE9-NEXT: xssubsp 0, 0, 3 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 2 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: addis 3, 2, .LCPI22_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI22_3@toc@l +; PC64LE9-NEXT: lxvx 36, 0, 3 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 3, 2, 4 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fsub_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI22_0@toc@ha +; PC64-NEXT: lfs 0, .LCPI22_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI22_1@toc@ha +; PC64-NEXT: lfs 3, .LCPI22_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI22_2@toc@ha +; PC64-NEXT: lfs 2, .LCPI22_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI22_3@toc@ha +; PC64-NEXT: lfs 4, .LCPI22_3@toc@l(3) +; PC64-NEXT: fsubs 1, 3, 0 +; PC64-NEXT: fsubs 2, 3, 2 +; PC64-NEXT: fsubs 3, 3, 4 +; PC64-NEXT: blr +entry: + %sub = call <3 x float> @llvm.experimental.constrained.fsub.v3f32( + <3 x float> , + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %sub +} + +define <3 x double> @constrained_vector_fsub_v3f64() { +; PC64LE-LABEL: constrained_vector_fsub_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI23_1@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI23_2@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI23_1@toc@l +; PC64LE-NEXT: addi 4, 4, .LCPI23_2@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: lxvd2x 1, 0, 4 +; PC64LE-NEXT: addis 3, 2, .LCPI23_0@toc@ha +; PC64LE-NEXT: lfd 3, .LCPI23_0@toc@l(3) +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xvsubdp 2, 1, 0 +; PC64LE-NEXT: xxlxor 0, 0, 0 +; PC64LE-NEXT: xssubdp 3, 3, 0 +; PC64LE-NEXT: xxswapd 1, 2 +; PC64LE-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fsub_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI23_0@toc@ha +; PC64LE9-NEXT: lfd 0, .LCPI23_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI23_1@toc@ha +; PC64LE9-NEXT: xxlxor 1, 1, 1 +; PC64LE9-NEXT: addi 3, 3, .LCPI23_1@toc@l +; PC64LE9-NEXT: xssubdp 3, 0, 1 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI23_2@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI23_2@toc@l +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: xvsubdp 2, 1, 0 +; PC64LE9-NEXT: xxswapd 1, 2 +; PC64LE9-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE9-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fsub_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI23_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI23_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI23_1@toc@ha +; PC64-NEXT: lfs 1, .LCPI23_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI23_2@toc@ha +; PC64-NEXT: lfs 2, .LCPI23_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI23_3@toc@ha +; PC64-NEXT: lfs 3, .LCPI23_3@toc@l(3) +; PC64-NEXT: fsub 1, 0, 1 +; PC64-NEXT: fsub 2, 0, 2 +; PC64-NEXT: fsub 3, 0, 3 +; PC64-NEXT: blr +entry: + %sub = call <3 x double> @llvm.experimental.constrained.fsub.v3f64( + <3 x double> , + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %sub +} + +define <4 x double> @constrained_vector_fsub_v4f64() { +; PC64LE-LABEL: constrained_vector_fsub_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI24_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI24_1@toc@ha +; PC64LE-NEXT: addis 5, 2, .LCPI24_2@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI24_0@toc@l +; PC64LE-NEXT: addi 4, 4, .LCPI24_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: addi 3, 5, .LCPI24_2@toc@l +; PC64LE-NEXT: lxvd2x 1, 0, 4 +; PC64LE-NEXT: lxvd2x 2, 0, 3 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xxswapd 2, 2 +; PC64LE-NEXT: xvsubdp 34, 1, 0 +; PC64LE-NEXT: xvsubdp 35, 1, 2 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fsub_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI24_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI24_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI24_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI24_1@toc@l +; PC64LE9-NEXT: lxvx 1, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI24_2@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI24_2@toc@l +; PC64LE9-NEXT: xvsubdp 34, 1, 0 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvsubdp 35, 1, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fsub_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI24_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI24_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI24_1@toc@ha +; PC64-NEXT: lfd 3, .LCPI24_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI24_2@toc@ha +; PC64-NEXT: lfd 1, .LCPI24_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI24_3@toc@ha +; PC64-NEXT: lfs 5, .LCPI24_3@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI24_4@toc@ha +; PC64-NEXT: fsub 2, 3, 0 +; PC64-NEXT: lfs 6, .LCPI24_4@toc@l(3) +; PC64-NEXT: fsub 4, 3, 1 +; PC64-NEXT: fsub 1, 3, 5 +; PC64-NEXT: fsub 3, 3, 6 +; PC64-NEXT: blr +entry: + %sub = call <4 x double> @llvm.experimental.constrained.fsub.v4f64( + <4 x double> , + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %sub +} + +define <1 x float> @constrained_vector_sqrt_v1f32() { +; PC64LE-LABEL: constrained_vector_sqrt_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI25_0@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI25_0@toc@l(3) +; PC64LE-NEXT: xssqrtsp 0, 0 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_sqrt_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI25_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI25_0@toc@l(3) +; PC64LE9-NEXT: xssqrtsp 0, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_sqrt_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI25_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI25_0@toc@l(3) +; PC64-NEXT: bl sqrtf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %sqrt = call <1 x float> @llvm.experimental.constrained.sqrt.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %sqrt +} + +define <2 x double> @constrained_vector_sqrt_v2f64() { +; PC64LE-LABEL: constrained_vector_sqrt_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI26_0@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI26_0@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xvsqrtdp 34, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_sqrt_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI26_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI26_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvsqrtdp 34, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_sqrt_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI26_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI26_0@toc@l(3) +; PC64-NEXT: bl sqrt +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI26_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI26_1@toc@l(3) +; PC64-NEXT: bl sqrt +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %sqrt = call <2 x double> @llvm.experimental.constrained.sqrt.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %sqrt +} + +define <3 x float> @constrained_vector_sqrt_v3f32() { +; PC64LE-LABEL: constrained_vector_sqrt_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI27_2@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI27_1@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI27_2@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI27_1@toc@l(4) +; PC64LE-NEXT: addis 3, 2, .LCPI27_0@toc@ha +; PC64LE-NEXT: xssqrtsp 0, 0 +; PC64LE-NEXT: lfs 2, .LCPI27_0@toc@l(3) +; PC64LE-NEXT: addis 3, 2, .LCPI27_3@toc@ha +; PC64LE-NEXT: xssqrtsp 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI27_3@toc@l +; PC64LE-NEXT: xssqrtsp 2, 2 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 2 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 3, 2 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_sqrt_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI27_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI27_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI27_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI27_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI27_2@toc@ha +; PC64LE9-NEXT: xssqrtsp 0, 0 +; PC64LE9-NEXT: lfs 2, .LCPI27_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI27_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI27_3@toc@l +; PC64LE9-NEXT: xssqrtsp 1, 1 +; PC64LE9-NEXT: xssqrtsp 2, 2 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xscvdpspn 1, 1 +; PC64LE9-NEXT: xscvdpspn 2, 2 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE9-NEXT: xxsldwi 34, 2, 2, 1 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_sqrt_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI27_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI27_0@toc@l(3) +; PC64-NEXT: bl sqrtf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI27_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI27_1@toc@l(3) +; PC64-NEXT: bl sqrtf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI27_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI27_2@toc@l(3) +; PC64-NEXT: bl sqrtf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %sqrt = call <3 x float> @llvm.experimental.constrained.sqrt.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %sqrt +} + +define <3 x double> @constrained_vector_sqrt_v3f64() { +; PC64LE-LABEL: constrained_vector_sqrt_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI28_1@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI28_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: addis 3, 2, .LCPI28_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI28_0@toc@l(3) +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xssqrtdp 3, 1 +; PC64LE-NEXT: xvsqrtdp 2, 0 +; PC64LE-NEXT: xxswapd 1, 2 +; PC64LE-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_sqrt_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI28_0@toc@ha +; PC64LE9-NEXT: lfd 0, .LCPI28_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI28_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI28_1@toc@l +; PC64LE9-NEXT: xssqrtdp 3, 0 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvsqrtdp 2, 0 +; PC64LE9-NEXT: xxswapd 1, 2 +; PC64LE9-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE9-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_sqrt_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI28_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI28_0@toc@l(3) +; PC64-NEXT: bl sqrt +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI28_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI28_1@toc@l(3) +; PC64-NEXT: bl sqrt +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI28_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI28_2@toc@l(3) +; PC64-NEXT: bl sqrt +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %sqrt = call <3 x double> @llvm.experimental.constrained.sqrt.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %sqrt +} + +define <4 x double> @constrained_vector_sqrt_v4f64() { +; PC64LE-LABEL: constrained_vector_sqrt_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI29_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI29_1@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI29_0@toc@l +; PC64LE-NEXT: addi 4, 4, .LCPI29_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: lxvd2x 1, 0, 4 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xvsqrtdp 34, 0 +; PC64LE-NEXT: xvsqrtdp 35, 1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_sqrt_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI29_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI29_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI29_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI29_1@toc@l +; PC64LE9-NEXT: xvsqrtdp 34, 0 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvsqrtdp 35, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_sqrt_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI29_0@toc@ha +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI29_0@toc@l(3) +; PC64-NEXT: bl sqrt +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI29_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI29_1@toc@l(3) +; PC64-NEXT: bl sqrt +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI29_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI29_2@toc@l(3) +; PC64-NEXT: bl sqrt +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI29_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI29_3@toc@l(3) +; PC64-NEXT: bl sqrt +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr + entry: + %sqrt = call <4 x double> @llvm.experimental.constrained.sqrt.v4f64( + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %sqrt +} + +define <1 x float> @constrained_vector_pow_v1f32() { +; PC64LE-LABEL: constrained_vector_pow_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI30_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI30_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI30_0@toc@l(3) +; PC64LE-NEXT: lfs 2, .LCPI30_1@toc@l(4) +; PC64LE-NEXT: bl powf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_pow_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI30_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI30_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI30_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI30_1@toc@l(3) +; PC64LE9-NEXT: bl powf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_pow_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI30_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI30_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI30_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI30_1@toc@l(3) +; PC64-NEXT: bl powf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %pow = call <1 x float> @llvm.experimental.constrained.pow.v1f32( + <1 x float> , + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %pow +} + +define <2 x double> @constrained_vector_pow_v2f64() { +; PC64LE-LABEL: constrained_vector_pow_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: addis 4, 2, .LCPI31_1@toc@ha +; PC64LE-NEXT: addis 3, 2, .LCPI31_0@toc@ha +; PC64LE-NEXT: lfs 31, .LCPI31_1@toc@l(4) +; PC64LE-NEXT: lfd 1, .LCPI31_0@toc@l(3) +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: bl pow +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI31_2@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI31_2@toc@l(3) +; PC64LE-NEXT: bl pow +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 1, 0 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_pow_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI31_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI31_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI31_1@toc@ha +; PC64LE9-NEXT: lfs 31, .LCPI31_1@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl pow +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI31_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI31_2@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl pow +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_pow_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI31_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI31_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI31_1@toc@ha +; PC64-NEXT: lfs 31, .LCPI31_1@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl pow +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI31_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI31_2@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl pow +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %pow = call <2 x double> @llvm.experimental.constrained.pow.v2f64( + <2 x double> , + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %pow +} + +define <3 x float> @constrained_vector_pow_v3f32() { +; PC64LE-LABEL: constrained_vector_pow_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f29, -24 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 29, -24(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: addis 4, 2, .LCPI32_1@toc@ha +; PC64LE-NEXT: addis 3, 2, .LCPI32_0@toc@ha +; PC64LE-NEXT: lfs 31, .LCPI32_1@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI32_0@toc@l(3) +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: bl powf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI32_2@toc@ha +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI32_2@toc@l(3) +; PC64LE-NEXT: bl powf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI32_3@toc@ha +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: fmr 29, 1 +; PC64LE-NEXT: lfs 1, .LCPI32_3@toc@l(3) +; PC64LE-NEXT: bl powf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 29 +; PC64LE-NEXT: addis 3, 2, .LCPI32_4@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI32_4@toc@l +; PC64LE-NEXT: lvx 4, 0, 3 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 3, 2, 4 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 29, -24(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_pow_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f29, -24 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 29, -24(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI32_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI32_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI32_1@toc@ha +; PC64LE9-NEXT: lfs 31, .LCPI32_1@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl powf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI32_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI32_2@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl powf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI32_3@toc@ha +; PC64LE9-NEXT: fmr 29, 1 +; PC64LE9-NEXT: lfs 1, .LCPI32_3@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl powf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 29 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: addis 3, 2, .LCPI32_4@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI32_4@toc@l +; PC64LE9-NEXT: lxvx 36, 0, 3 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 3, 2, 4 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 29, -24(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_pow_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI32_0@toc@ha +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI32_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI32_1@toc@ha +; PC64-NEXT: lfs 31, .LCPI32_1@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl powf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI32_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI32_2@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl powf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI32_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfs 1, .LCPI32_3@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl powf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 30 +; PC64-NEXT: fmr 2, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %pow = call <3 x float> @llvm.experimental.constrained.pow.v3f32( + <3 x float> , + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %pow +} + +define <3 x double> @constrained_vector_pow_v3f64() { +; PC64LE-LABEL: constrained_vector_pow_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -96(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 96 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: .cfi_offset v31, -32 +; PC64LE-NEXT: addis 4, 2, .LCPI33_1@toc@ha +; PC64LE-NEXT: stfd 31, 88(1) # 8-byte Folded Spill +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI33_0@toc@ha +; PC64LE-NEXT: lfs 31, .LCPI33_1@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI33_0@toc@l(3) +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: bl pow +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI33_2@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI33_2@toc@l(3) +; PC64LE-NEXT: bl pow +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI33_3@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfd 1, .LCPI33_3@toc@l(3) +; PC64LE-NEXT: bl pow +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: lfd 31, 88(1) # 8-byte Folded Reload +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 96 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_pow_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -80(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 80 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: .cfi_offset v31, -32 +; PC64LE9-NEXT: addis 3, 2, .LCPI33_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI33_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI33_1@toc@ha +; PC64LE9-NEXT: stfd 31, 72(1) # 8-byte Folded Spill +; PC64LE9-NEXT: lfs 31, .LCPI33_1@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl pow +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI33_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI33_2@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl pow +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI33_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfd 1, .LCPI33_3@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl pow +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: lfd 31, 72(1) # 8-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 80 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_pow_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI33_0@toc@ha +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI33_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI33_1@toc@ha +; PC64-NEXT: lfs 31, .LCPI33_1@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl pow +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI33_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI33_2@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl pow +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI33_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI33_3@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl pow +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 30 +; PC64-NEXT: fmr 2, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %pow = call <3 x double> @llvm.experimental.constrained.pow.v3f64( + <3 x double> , + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %pow +} + +define <4 x double> @constrained_vector_pow_v4f64() { +; PC64LE-LABEL: constrained_vector_pow_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -96(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 96 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: .cfi_offset v31, -32 +; PC64LE-NEXT: addis 4, 2, .LCPI34_1@toc@ha +; PC64LE-NEXT: stfd 31, 88(1) # 8-byte Folded Spill +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI34_0@toc@ha +; PC64LE-NEXT: lfs 31, .LCPI34_1@toc@l(4) +; PC64LE-NEXT: lfd 1, .LCPI34_0@toc@l(3) +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: bl pow +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI34_2@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI34_2@toc@l(3) +; PC64LE-NEXT: bl pow +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI34_3@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfd 1, .LCPI34_3@toc@l(3) +; PC64LE-NEXT: bl pow +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI34_4@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI34_4@toc@l(3) +; PC64LE-NEXT: bl pow +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 2, 31 +; PC64LE-NEXT: lfd 31, 88(1) # 8-byte Folded Reload +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 35, 1, 0 +; PC64LE-NEXT: addi 1, 1, 96 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_pow_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -80(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 80 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: .cfi_offset v31, -32 +; PC64LE9-NEXT: addis 3, 2, .LCPI34_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI34_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI34_1@toc@ha +; PC64LE9-NEXT: stfd 31, 72(1) # 8-byte Folded Spill +; PC64LE9-NEXT: lfs 31, .LCPI34_1@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl pow +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI34_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI34_2@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl pow +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI34_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfd 1, .LCPI34_3@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl pow +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI34_4@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI34_4@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl pow +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 2, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: lfd 31, 72(1) # 8-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 35, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 80 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_pow_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f28, -32 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI34_0@toc@ha +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 28, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI34_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI34_1@toc@ha +; PC64-NEXT: lfs 31, .LCPI34_1@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl pow +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI34_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI34_2@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl pow +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI34_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI34_3@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl pow +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI34_4@toc@ha +; PC64-NEXT: fmr 28, 1 +; PC64-NEXT: lfd 1, .LCPI34_4@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl pow +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 30 +; PC64-NEXT: fmr 2, 29 +; PC64-NEXT: fmr 3, 28 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 28, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %pow = call <4 x double> @llvm.experimental.constrained.pow.v4f64( + <4 x double> , + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %pow +} + +define <1 x float> @constrained_vector_powi_v1f32() { +; PC64LE-LABEL: constrained_vector_powi_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI35_0@toc@ha +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: lfs 1, .LCPI35_0@toc@l(3) +; PC64LE-NEXT: bl __powisf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_powi_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI35_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI35_0@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: bl __powisf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_powi_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI35_0@toc@ha +; PC64-NEXT: li 4, 3 +; PC64-NEXT: lfs 1, .LCPI35_0@toc@l(3) +; PC64-NEXT: bl __powisf2 +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %powi = call <1 x float> @llvm.experimental.constrained.powi.v1f32( + <1 x float> , + i32 3, + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %powi +} + +define <2 x double> @constrained_vector_powi_v2f64() { +; PC64LE-LABEL: constrained_vector_powi_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI36_0@toc@ha +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: lfd 1, .LCPI36_0@toc@l(3) +; PC64LE-NEXT: bl __powidf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI36_1@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI36_1@toc@l(3) +; PC64LE-NEXT: bl __powidf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 1, 0 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_powi_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI36_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI36_0@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: bl __powidf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI36_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI36_1@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: bl __powidf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_powi_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI36_0@toc@ha +; PC64-NEXT: li 4, 3 +; PC64-NEXT: lfd 1, .LCPI36_0@toc@l(3) +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: bl __powidf2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI36_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI36_1@toc@l(3) +; PC64-NEXT: li 4, 3 +; PC64-NEXT: bl __powidf2 +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %powi = call <2 x double> @llvm.experimental.constrained.powi.v2f64( + <2 x double> , + i32 3, + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %powi +} + +define <3 x float> @constrained_vector_powi_v3f32() { +; +; +; PC64LE-LABEL: constrained_vector_powi_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 48 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -48(1) +; PC64LE-NEXT: addis 3, 2, .LCPI37_0@toc@ha +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: lfs 1, .LCPI37_0@toc@l(3) +; PC64LE-NEXT: bl __powisf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI37_1@toc@ha +; PC64LE-NEXT: fmr 31, 1 +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: lfs 1, .LCPI37_1@toc@l(3) +; PC64LE-NEXT: bl __powisf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI37_2@toc@ha +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: lfs 1, .LCPI37_2@toc@l(3) +; PC64LE-NEXT: bl __powisf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: addis 3, 2, .LCPI37_3@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI37_3@toc@l +; PC64LE-NEXT: lvx 4, 0, 3 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 31 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 3, 2, 4 +; PC64LE-NEXT: addi 1, 1, 48 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_powi_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI37_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI37_0@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: bl __powisf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI37_1@toc@ha +; PC64LE9-NEXT: fmr 31, 1 +; PC64LE9-NEXT: lfs 1, .LCPI37_1@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: bl __powisf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI37_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI37_2@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: bl __powisf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 31 +; PC64LE9-NEXT: addis 3, 2, .LCPI37_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI37_3@toc@l +; PC64LE9-NEXT: lxvx 36, 0, 3 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 3, 2, 4 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_powi_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI37_0@toc@ha +; PC64-NEXT: li 4, 3 +; PC64-NEXT: lfs 1, .LCPI37_0@toc@l(3) +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: bl __powisf2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI37_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI37_1@toc@l(3) +; PC64-NEXT: li 4, 3 +; PC64-NEXT: bl __powisf2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI37_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI37_2@toc@l(3) +; PC64-NEXT: li 4, 3 +; PC64-NEXT: bl __powisf2 +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %powi = call <3 x float> @llvm.experimental.constrained.powi.v3f32( + <3 x float> , + i32 3, + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %powi +} + +define <3 x double> @constrained_vector_powi_v3f64() { +; PC64LE-LABEL: constrained_vector_powi_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI38_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI38_0@toc@l(3) +; PC64LE-NEXT: bl __powidf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI38_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI38_1@toc@l(3) +; PC64LE-NEXT: bl __powidf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI38_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 0, 1 +; PC64LE-NEXT: lfd 1, .LCPI38_2@toc@l(3) +; PC64LE-NEXT: bl __powidf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_powi_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI38_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI38_0@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl __powidf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI38_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI38_1@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: bl __powidf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI38_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 0, 1 +; PC64LE9-NEXT: lfd 1, .LCPI38_2@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: bl __powidf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_powi_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI38_0@toc@ha +; PC64-NEXT: li 4, 3 +; PC64-NEXT: lfs 1, .LCPI38_0@toc@l(3) +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: bl __powidf2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI38_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI38_1@toc@l(3) +; PC64-NEXT: li 4, 3 +; PC64-NEXT: bl __powidf2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI38_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI38_2@toc@l(3) +; PC64-NEXT: li 4, 3 +; PC64-NEXT: bl __powidf2 +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %powi = call <3 x double> @llvm.experimental.constrained.powi.v3f64( + <3 x double> , + i32 3, + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %powi +} + +define <4 x double> @constrained_vector_powi_v4f64() { +; PC64LE-LABEL: constrained_vector_powi_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI39_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI39_0@toc@l(3) +; PC64LE-NEXT: bl __powidf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI39_1@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI39_1@toc@l(3) +; PC64LE-NEXT: bl __powidf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI39_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfd 1, .LCPI39_2@toc@l(3) +; PC64LE-NEXT: bl __powidf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: li 4, 3 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI39_3@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI39_3@toc@l(3) +; PC64LE-NEXT: bl __powidf2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 2, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 35, 1, 0 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_powi_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI39_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI39_0@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl __powidf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI39_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI39_1@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: bl __powidf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI39_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfd 1, .LCPI39_2@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: bl __powidf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI39_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI39_3@toc@l(3) +; PC64LE9-NEXT: li 4, 3 +; PC64LE9-NEXT: bl __powidf2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 2, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 35, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_powi_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI39_0@toc@ha +; PC64-NEXT: li 4, 3 +; PC64-NEXT: lfd 1, .LCPI39_0@toc@l(3) +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: bl __powidf2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI39_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI39_1@toc@l(3) +; PC64-NEXT: li 4, 3 +; PC64-NEXT: bl __powidf2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI39_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI39_2@toc@l(3) +; PC64-NEXT: li 4, 3 +; PC64-NEXT: bl __powidf2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI39_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI39_3@toc@l(3) +; PC64-NEXT: li 4, 3 +; PC64-NEXT: bl __powidf2 +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %powi = call <4 x double> @llvm.experimental.constrained.powi.v4f64( + <4 x double> , + i32 3, + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %powi +} + +define <1 x float> @constrained_vector_sin_v1f32() { +; PC64LE-LABEL: constrained_vector_sin_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI40_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI40_0@toc@l(3) +; PC64LE-NEXT: bl sinf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_sin_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI40_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI40_0@toc@l(3) +; PC64LE9-NEXT: bl sinf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_sin_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI40_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI40_0@toc@l(3) +; PC64-NEXT: bl sinf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %sin = call <1 x float> @llvm.experimental.constrained.sin.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %sin +} + +define <2 x double> @constrained_vector_sin_v2f64() { +; PC64LE-LABEL: constrained_vector_sin_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI41_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI41_0@toc@l(3) +; PC64LE-NEXT: bl sin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI41_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI41_1@toc@l(3) +; PC64LE-NEXT: bl sin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 0, 1 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_sin_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI41_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI41_0@toc@l(3) +; PC64LE9-NEXT: bl sin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI41_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI41_1@toc@l(3) +; PC64LE9-NEXT: bl sin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_sin_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI41_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI41_0@toc@l(3) +; PC64-NEXT: bl sin +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI41_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI41_1@toc@l(3) +; PC64-NEXT: bl sin +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %sin = call <2 x double> @llvm.experimental.constrained.sin.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %sin +} + +define <3 x float> @constrained_vector_sin_v3f32() { +; PC64LE-LABEL: constrained_vector_sin_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 48 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -48(1) +; PC64LE-NEXT: addis 3, 2, .LCPI42_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI42_0@toc@l(3) +; PC64LE-NEXT: bl sinf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI42_1@toc@ha +; PC64LE-NEXT: fmr 31, 1 +; PC64LE-NEXT: lfs 1, .LCPI42_1@toc@l(3) +; PC64LE-NEXT: bl sinf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI42_2@toc@ha +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI42_2@toc@l(3) +; PC64LE-NEXT: bl sinf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: addis 3, 2, .LCPI42_3@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI42_3@toc@l +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 31 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: addi 1, 1, 48 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_sin_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI42_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI42_0@toc@l(3) +; PC64LE9-NEXT: bl sinf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI42_1@toc@ha +; PC64LE9-NEXT: fmr 31, 1 +; PC64LE9-NEXT: lfs 1, .LCPI42_1@toc@l(3) +; PC64LE9-NEXT: bl sinf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI42_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI42_2@toc@l(3) +; PC64LE9-NEXT: bl sinf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 31 +; PC64LE9-NEXT: addis 3, 2, .LCPI42_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI42_3@toc@l +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_sin_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI42_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI42_0@toc@l(3) +; PC64-NEXT: bl sinf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI42_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI42_1@toc@l(3) +; PC64-NEXT: bl sinf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI42_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI42_2@toc@l(3) +; PC64-NEXT: bl sinf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %sin = call <3 x float> @llvm.experimental.constrained.sin.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %sin +} + +define <3 x double> @constrained_vector_sin_v3f64() { +; PC64LE-LABEL: constrained_vector_sin_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI43_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI43_0@toc@l(3) +; PC64LE-NEXT: bl sin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI43_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI43_1@toc@l(3) +; PC64LE-NEXT: bl sin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI43_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 0, 1 +; PC64LE-NEXT: lfd 1, .LCPI43_2@toc@l(3) +; PC64LE-NEXT: bl sin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_sin_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI43_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI43_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl sin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI43_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI43_1@toc@l(3) +; PC64LE9-NEXT: bl sin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI43_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 0, 1 +; PC64LE9-NEXT: lfd 1, .LCPI43_2@toc@l(3) +; PC64LE9-NEXT: bl sin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_sin_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI43_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI43_0@toc@l(3) +; PC64-NEXT: bl sin +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI43_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI43_1@toc@l(3) +; PC64-NEXT: bl sin +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI43_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI43_2@toc@l(3) +; PC64-NEXT: bl sin +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %sin = call <3 x double> @llvm.experimental.constrained.sin.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %sin +} + +define <4 x double> @constrained_vector_sin_v4f64() { +; PC64LE-LABEL: constrained_vector_sin_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI44_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI44_0@toc@l(3) +; PC64LE-NEXT: bl sin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI44_1@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI44_1@toc@l(3) +; PC64LE-NEXT: bl sin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI44_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfd 1, .LCPI44_2@toc@l(3) +; PC64LE-NEXT: bl sin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI44_3@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI44_3@toc@l(3) +; PC64LE-NEXT: bl sin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 3, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 0, 1 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_sin_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI44_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI44_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl sin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI44_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI44_1@toc@l(3) +; PC64LE9-NEXT: bl sin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI44_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfd 1, .LCPI44_2@toc@l(3) +; PC64LE9-NEXT: bl sin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI44_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI44_3@toc@l(3) +; PC64LE9-NEXT: bl sin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 3, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_sin_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI44_0@toc@ha +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI44_0@toc@l(3) +; PC64-NEXT: bl sin +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI44_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI44_1@toc@l(3) +; PC64-NEXT: bl sin +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI44_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI44_2@toc@l(3) +; PC64-NEXT: bl sin +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI44_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI44_3@toc@l(3) +; PC64-NEXT: bl sin +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %sin = call <4 x double> @llvm.experimental.constrained.sin.v4f64( + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %sin +} + +define <1 x float> @constrained_vector_cos_v1f32() { +; PC64LE-LABEL: constrained_vector_cos_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI45_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI45_0@toc@l(3) +; PC64LE-NEXT: bl cosf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_cos_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI45_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI45_0@toc@l(3) +; PC64LE9-NEXT: bl cosf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_cos_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI45_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI45_0@toc@l(3) +; PC64-NEXT: bl cosf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %cos = call <1 x float> @llvm.experimental.constrained.cos.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %cos +} + +define <2 x double> @constrained_vector_cos_v2f64() { +; PC64LE-LABEL: constrained_vector_cos_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI46_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI46_0@toc@l(3) +; PC64LE-NEXT: bl cos +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI46_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI46_1@toc@l(3) +; PC64LE-NEXT: bl cos +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 0, 1 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_cos_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI46_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI46_0@toc@l(3) +; PC64LE9-NEXT: bl cos +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI46_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI46_1@toc@l(3) +; PC64LE9-NEXT: bl cos +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_cos_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI46_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI46_0@toc@l(3) +; PC64-NEXT: bl cos +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI46_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI46_1@toc@l(3) +; PC64-NEXT: bl cos +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %cos = call <2 x double> @llvm.experimental.constrained.cos.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %cos +} + +define <3 x float> @constrained_vector_cos_v3f32() { +; PC64LE-LABEL: constrained_vector_cos_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 48 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -48(1) +; PC64LE-NEXT: addis 3, 2, .LCPI47_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI47_0@toc@l(3) +; PC64LE-NEXT: bl cosf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI47_1@toc@ha +; PC64LE-NEXT: fmr 31, 1 +; PC64LE-NEXT: lfs 1, .LCPI47_1@toc@l(3) +; PC64LE-NEXT: bl cosf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI47_2@toc@ha +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI47_2@toc@l(3) +; PC64LE-NEXT: bl cosf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: addis 3, 2, .LCPI47_3@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI47_3@toc@l +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 31 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: addi 1, 1, 48 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_cos_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI47_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI47_0@toc@l(3) +; PC64LE9-NEXT: bl cosf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI47_1@toc@ha +; PC64LE9-NEXT: fmr 31, 1 +; PC64LE9-NEXT: lfs 1, .LCPI47_1@toc@l(3) +; PC64LE9-NEXT: bl cosf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI47_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI47_2@toc@l(3) +; PC64LE9-NEXT: bl cosf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 31 +; PC64LE9-NEXT: addis 3, 2, .LCPI47_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI47_3@toc@l +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_cos_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI47_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI47_0@toc@l(3) +; PC64-NEXT: bl cosf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI47_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI47_1@toc@l(3) +; PC64-NEXT: bl cosf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI47_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI47_2@toc@l(3) +; PC64-NEXT: bl cosf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %cos = call <3 x float> @llvm.experimental.constrained.cos.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %cos +} + +define <3 x double> @constrained_vector_cos_v3f64() { +; PC64LE-LABEL: constrained_vector_cos_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI48_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI48_0@toc@l(3) +; PC64LE-NEXT: bl cos +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI48_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI48_1@toc@l(3) +; PC64LE-NEXT: bl cos +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI48_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 0, 1 +; PC64LE-NEXT: lfd 1, .LCPI48_2@toc@l(3) +; PC64LE-NEXT: bl cos +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_cos_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI48_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI48_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl cos +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI48_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI48_1@toc@l(3) +; PC64LE9-NEXT: bl cos +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI48_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 0, 1 +; PC64LE9-NEXT: lfd 1, .LCPI48_2@toc@l(3) +; PC64LE9-NEXT: bl cos +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_cos_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI48_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI48_0@toc@l(3) +; PC64-NEXT: bl cos +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI48_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI48_1@toc@l(3) +; PC64-NEXT: bl cos +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI48_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI48_2@toc@l(3) +; PC64-NEXT: bl cos +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %cos = call <3 x double> @llvm.experimental.constrained.cos.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %cos +} + +define <4 x double> @constrained_vector_cos_v4f64() { +; PC64LE-LABEL: constrained_vector_cos_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI49_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI49_0@toc@l(3) +; PC64LE-NEXT: bl cos +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI49_1@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI49_1@toc@l(3) +; PC64LE-NEXT: bl cos +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI49_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfd 1, .LCPI49_2@toc@l(3) +; PC64LE-NEXT: bl cos +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI49_3@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI49_3@toc@l(3) +; PC64LE-NEXT: bl cos +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 3, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 0, 1 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_cos_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI49_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI49_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl cos +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI49_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI49_1@toc@l(3) +; PC64LE9-NEXT: bl cos +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI49_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfd 1, .LCPI49_2@toc@l(3) +; PC64LE9-NEXT: bl cos +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI49_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI49_3@toc@l(3) +; PC64LE9-NEXT: bl cos +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 3, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_cos_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI49_0@toc@ha +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI49_0@toc@l(3) +; PC64-NEXT: bl cos +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI49_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI49_1@toc@l(3) +; PC64-NEXT: bl cos +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI49_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI49_2@toc@l(3) +; PC64-NEXT: bl cos +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI49_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI49_3@toc@l(3) +; PC64-NEXT: bl cos +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %cos = call <4 x double> @llvm.experimental.constrained.cos.v4f64( + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %cos +} + +define <1 x float> @constrained_vector_exp_v1f32() { +; PC64LE-LABEL: constrained_vector_exp_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI50_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI50_0@toc@l(3) +; PC64LE-NEXT: bl expf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_exp_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI50_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI50_0@toc@l(3) +; PC64LE9-NEXT: bl expf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_exp_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI50_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI50_0@toc@l(3) +; PC64-NEXT: bl expf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %exp = call <1 x float> @llvm.experimental.constrained.exp.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %exp +} + +define <2 x double> @constrained_vector_exp_v2f64() { +; PC64LE-LABEL: constrained_vector_exp_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI51_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI51_0@toc@l(3) +; PC64LE-NEXT: bl exp +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI51_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI51_1@toc@l(3) +; PC64LE-NEXT: bl exp +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 0, 1 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_exp_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI51_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI51_0@toc@l(3) +; PC64LE9-NEXT: bl exp +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI51_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI51_1@toc@l(3) +; PC64LE9-NEXT: bl exp +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_exp_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI51_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI51_0@toc@l(3) +; PC64-NEXT: bl exp +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI51_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI51_1@toc@l(3) +; PC64-NEXT: bl exp +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %exp = call <2 x double> @llvm.experimental.constrained.exp.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %exp +} + +define <3 x float> @constrained_vector_exp_v3f32() { +; PC64LE-LABEL: constrained_vector_exp_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 48 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -48(1) +; PC64LE-NEXT: addis 3, 2, .LCPI52_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI52_0@toc@l(3) +; PC64LE-NEXT: bl expf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI52_1@toc@ha +; PC64LE-NEXT: fmr 31, 1 +; PC64LE-NEXT: lfs 1, .LCPI52_1@toc@l(3) +; PC64LE-NEXT: bl expf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI52_2@toc@ha +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI52_2@toc@l(3) +; PC64LE-NEXT: bl expf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: addis 3, 2, .LCPI52_3@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI52_3@toc@l +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 31 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: addi 1, 1, 48 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_exp_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI52_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI52_0@toc@l(3) +; PC64LE9-NEXT: bl expf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI52_1@toc@ha +; PC64LE9-NEXT: fmr 31, 1 +; PC64LE9-NEXT: lfs 1, .LCPI52_1@toc@l(3) +; PC64LE9-NEXT: bl expf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI52_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI52_2@toc@l(3) +; PC64LE9-NEXT: bl expf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 31 +; PC64LE9-NEXT: addis 3, 2, .LCPI52_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI52_3@toc@l +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_exp_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI52_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI52_0@toc@l(3) +; PC64-NEXT: bl expf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI52_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI52_1@toc@l(3) +; PC64-NEXT: bl expf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI52_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI52_2@toc@l(3) +; PC64-NEXT: bl expf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %exp = call <3 x float> @llvm.experimental.constrained.exp.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %exp +} + +define <3 x double> @constrained_vector_exp_v3f64() { +; PC64LE-LABEL: constrained_vector_exp_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI53_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI53_0@toc@l(3) +; PC64LE-NEXT: bl exp +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI53_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI53_1@toc@l(3) +; PC64LE-NEXT: bl exp +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI53_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 0, 1 +; PC64LE-NEXT: lfd 1, .LCPI53_2@toc@l(3) +; PC64LE-NEXT: bl exp +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_exp_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI53_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI53_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl exp +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI53_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI53_1@toc@l(3) +; PC64LE9-NEXT: bl exp +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI53_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 0, 1 +; PC64LE9-NEXT: lfd 1, .LCPI53_2@toc@l(3) +; PC64LE9-NEXT: bl exp +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_exp_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI53_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI53_0@toc@l(3) +; PC64-NEXT: bl exp +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI53_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI53_1@toc@l(3) +; PC64-NEXT: bl exp +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI53_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI53_2@toc@l(3) +; PC64-NEXT: bl exp +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %exp = call <3 x double> @llvm.experimental.constrained.exp.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %exp +} + +define <4 x double> @constrained_vector_exp_v4f64() { +; PC64LE-LABEL: constrained_vector_exp_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI54_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI54_0@toc@l(3) +; PC64LE-NEXT: bl exp +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI54_1@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI54_1@toc@l(3) +; PC64LE-NEXT: bl exp +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI54_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfd 1, .LCPI54_2@toc@l(3) +; PC64LE-NEXT: bl exp +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI54_3@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI54_3@toc@l(3) +; PC64LE-NEXT: bl exp +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 3, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 0, 1 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_exp_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI54_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI54_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl exp +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI54_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI54_1@toc@l(3) +; PC64LE9-NEXT: bl exp +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI54_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfd 1, .LCPI54_2@toc@l(3) +; PC64LE9-NEXT: bl exp +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI54_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI54_3@toc@l(3) +; PC64LE9-NEXT: bl exp +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 3, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_exp_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI54_0@toc@ha +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI54_0@toc@l(3) +; PC64-NEXT: bl exp +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI54_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI54_1@toc@l(3) +; PC64-NEXT: bl exp +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI54_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI54_2@toc@l(3) +; PC64-NEXT: bl exp +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI54_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI54_3@toc@l(3) +; PC64-NEXT: bl exp +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %exp = call <4 x double> @llvm.experimental.constrained.exp.v4f64( + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %exp +} + +define <1 x float> @constrained_vector_exp2_v1f32() { +; PC64LE-LABEL: constrained_vector_exp2_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI55_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI55_0@toc@l(3) +; PC64LE-NEXT: bl exp2f +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_exp2_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI55_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI55_0@toc@l(3) +; PC64LE9-NEXT: bl exp2f +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_exp2_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI55_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI55_0@toc@l(3) +; PC64-NEXT: bl exp2f +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %exp2 = call <1 x float> @llvm.experimental.constrained.exp2.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %exp2 +} + +define <2 x double> @constrained_vector_exp2_v2f64() { +; PC64LE-LABEL: constrained_vector_exp2_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI56_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI56_0@toc@l(3) +; PC64LE-NEXT: bl exp2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI56_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI56_1@toc@l(3) +; PC64LE-NEXT: bl exp2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 1, 0 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_exp2_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI56_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI56_0@toc@l(3) +; PC64LE9-NEXT: bl exp2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI56_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI56_1@toc@l(3) +; PC64LE9-NEXT: bl exp2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_exp2_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI56_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI56_0@toc@l(3) +; PC64-NEXT: bl exp2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI56_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI56_1@toc@l(3) +; PC64-NEXT: bl exp2 +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %exp2 = call <2 x double> @llvm.experimental.constrained.exp2.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %exp2 +} + +define <3 x float> @constrained_vector_exp2_v3f32() { +; PC64LE-LABEL: constrained_vector_exp2_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 48 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -48(1) +; PC64LE-NEXT: addis 3, 2, .LCPI57_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI57_0@toc@l(3) +; PC64LE-NEXT: bl exp2f +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI57_1@toc@ha +; PC64LE-NEXT: fmr 31, 1 +; PC64LE-NEXT: lfs 1, .LCPI57_1@toc@l(3) +; PC64LE-NEXT: bl exp2f +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI57_2@toc@ha +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI57_2@toc@l(3) +; PC64LE-NEXT: bl exp2f +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: addis 3, 2, .LCPI57_3@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI57_3@toc@l +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 31 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: addi 1, 1, 48 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_exp2_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI57_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI57_0@toc@l(3) +; PC64LE9-NEXT: bl exp2f +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI57_1@toc@ha +; PC64LE9-NEXT: fmr 31, 1 +; PC64LE9-NEXT: lfs 1, .LCPI57_1@toc@l(3) +; PC64LE9-NEXT: bl exp2f +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI57_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI57_2@toc@l(3) +; PC64LE9-NEXT: bl exp2f +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 31 +; PC64LE9-NEXT: addis 3, 2, .LCPI57_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI57_3@toc@l +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_exp2_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI57_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI57_0@toc@l(3) +; PC64-NEXT: bl exp2f +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI57_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI57_1@toc@l(3) +; PC64-NEXT: bl exp2f +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI57_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI57_2@toc@l(3) +; PC64-NEXT: bl exp2f +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %exp2 = call <3 x float> @llvm.experimental.constrained.exp2.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %exp2 +} + +define <3 x double> @constrained_vector_exp2_v3f64() { +; PC64LE-LABEL: constrained_vector_exp2_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI58_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI58_0@toc@l(3) +; PC64LE-NEXT: bl exp2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI58_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI58_1@toc@l(3) +; PC64LE-NEXT: bl exp2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI58_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 0, 1 +; PC64LE-NEXT: lfd 1, .LCPI58_2@toc@l(3) +; PC64LE-NEXT: bl exp2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_exp2_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI58_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI58_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl exp2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI58_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI58_1@toc@l(3) +; PC64LE9-NEXT: bl exp2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI58_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 0, 1 +; PC64LE9-NEXT: lfd 1, .LCPI58_2@toc@l(3) +; PC64LE9-NEXT: bl exp2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_exp2_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI58_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI58_0@toc@l(3) +; PC64-NEXT: bl exp2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI58_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI58_1@toc@l(3) +; PC64-NEXT: bl exp2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI58_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI58_2@toc@l(3) +; PC64-NEXT: bl exp2 +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %exp2 = call <3 x double> @llvm.experimental.constrained.exp2.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %exp2 +} + +define <4 x double> @constrained_vector_exp2_v4f64() { +; PC64LE-LABEL: constrained_vector_exp2_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI59_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI59_0@toc@l(3) +; PC64LE-NEXT: bl exp2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI59_1@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI59_1@toc@l(3) +; PC64LE-NEXT: bl exp2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI59_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfd 1, .LCPI59_2@toc@l(3) +; PC64LE-NEXT: bl exp2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI59_3@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI59_3@toc@l(3) +; PC64LE-NEXT: bl exp2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 2, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 35, 1, 0 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_exp2_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI59_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI59_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl exp2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI59_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI59_1@toc@l(3) +; PC64LE9-NEXT: bl exp2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI59_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfd 1, .LCPI59_2@toc@l(3) +; PC64LE9-NEXT: bl exp2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI59_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI59_3@toc@l(3) +; PC64LE9-NEXT: bl exp2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 2, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 35, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_exp2_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI59_0@toc@ha +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI59_0@toc@l(3) +; PC64-NEXT: bl exp2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI59_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI59_1@toc@l(3) +; PC64-NEXT: bl exp2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI59_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI59_2@toc@l(3) +; PC64-NEXT: bl exp2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI59_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI59_3@toc@l(3) +; PC64-NEXT: bl exp2 +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %exp2 = call <4 x double> @llvm.experimental.constrained.exp2.v4f64( + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %exp2 +} + +define <1 x float> @constrained_vector_log_v1f32() { +; PC64LE-LABEL: constrained_vector_log_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI60_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI60_0@toc@l(3) +; PC64LE-NEXT: bl logf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI60_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI60_0@toc@l(3) +; PC64LE9-NEXT: bl logf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI60_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI60_0@toc@l(3) +; PC64-NEXT: bl logf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log = call <1 x float> @llvm.experimental.constrained.log.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %log +} + +define <2 x double> @constrained_vector_log_v2f64() { +; PC64LE-LABEL: constrained_vector_log_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI61_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI61_0@toc@l(3) +; PC64LE-NEXT: bl log +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI61_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI61_1@toc@l(3) +; PC64LE-NEXT: bl log +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 0, 1 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI61_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI61_0@toc@l(3) +; PC64LE9-NEXT: bl log +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI61_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI61_1@toc@l(3) +; PC64LE9-NEXT: bl log +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI61_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI61_0@toc@l(3) +; PC64-NEXT: bl log +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI61_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI61_1@toc@l(3) +; PC64-NEXT: bl log +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log = call <2 x double> @llvm.experimental.constrained.log.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %log +} + +define <3 x float> @constrained_vector_log_v3f32() { +; PC64LE-LABEL: constrained_vector_log_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 48 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -48(1) +; PC64LE-NEXT: addis 3, 2, .LCPI62_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI62_0@toc@l(3) +; PC64LE-NEXT: bl logf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI62_1@toc@ha +; PC64LE-NEXT: fmr 31, 1 +; PC64LE-NEXT: lfs 1, .LCPI62_1@toc@l(3) +; PC64LE-NEXT: bl logf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI62_2@toc@ha +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI62_2@toc@l(3) +; PC64LE-NEXT: bl logf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: addis 3, 2, .LCPI62_3@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI62_3@toc@l +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 31 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: addi 1, 1, 48 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI62_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI62_0@toc@l(3) +; PC64LE9-NEXT: bl logf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI62_1@toc@ha +; PC64LE9-NEXT: fmr 31, 1 +; PC64LE9-NEXT: lfs 1, .LCPI62_1@toc@l(3) +; PC64LE9-NEXT: bl logf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI62_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI62_2@toc@l(3) +; PC64LE9-NEXT: bl logf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 31 +; PC64LE9-NEXT: addis 3, 2, .LCPI62_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI62_3@toc@l +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI62_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI62_0@toc@l(3) +; PC64-NEXT: bl logf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI62_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI62_1@toc@l(3) +; PC64-NEXT: bl logf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI62_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI62_2@toc@l(3) +; PC64-NEXT: bl logf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log = call <3 x float> @llvm.experimental.constrained.log.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %log +} + +define <3 x double> @constrained_vector_log_v3f64() { +; PC64LE-LABEL: constrained_vector_log_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI63_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI63_0@toc@l(3) +; PC64LE-NEXT: bl log +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI63_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI63_1@toc@l(3) +; PC64LE-NEXT: bl log +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI63_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 0, 1 +; PC64LE-NEXT: lfd 1, .LCPI63_2@toc@l(3) +; PC64LE-NEXT: bl log +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI63_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI63_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl log +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI63_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI63_1@toc@l(3) +; PC64LE9-NEXT: bl log +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI63_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 0, 1 +; PC64LE9-NEXT: lfd 1, .LCPI63_2@toc@l(3) +; PC64LE9-NEXT: bl log +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI63_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI63_0@toc@l(3) +; PC64-NEXT: bl log +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI63_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI63_1@toc@l(3) +; PC64-NEXT: bl log +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI63_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI63_2@toc@l(3) +; PC64-NEXT: bl log +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log = call <3 x double> @llvm.experimental.constrained.log.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %log +} + +define <4 x double> @constrained_vector_log_v4f64() { +; PC64LE-LABEL: constrained_vector_log_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI64_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI64_0@toc@l(3) +; PC64LE-NEXT: bl log +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI64_1@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI64_1@toc@l(3) +; PC64LE-NEXT: bl log +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI64_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfd 1, .LCPI64_2@toc@l(3) +; PC64LE-NEXT: bl log +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI64_3@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI64_3@toc@l(3) +; PC64LE-NEXT: bl log +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 3, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 0, 1 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI64_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI64_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl log +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI64_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI64_1@toc@l(3) +; PC64LE9-NEXT: bl log +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI64_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfd 1, .LCPI64_2@toc@l(3) +; PC64LE9-NEXT: bl log +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI64_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI64_3@toc@l(3) +; PC64LE9-NEXT: bl log +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 3, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI64_0@toc@ha +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI64_0@toc@l(3) +; PC64-NEXT: bl log +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI64_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI64_1@toc@l(3) +; PC64-NEXT: bl log +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI64_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI64_2@toc@l(3) +; PC64-NEXT: bl log +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI64_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI64_3@toc@l(3) +; PC64-NEXT: bl log +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log = call <4 x double> @llvm.experimental.constrained.log.v4f64( + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %log +} + +define <1 x float> @constrained_vector_log10_v1f32() { +; PC64LE-LABEL: constrained_vector_log10_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI65_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI65_0@toc@l(3) +; PC64LE-NEXT: bl log10f +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log10_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI65_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI65_0@toc@l(3) +; PC64LE9-NEXT: bl log10f +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log10_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI65_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI65_0@toc@l(3) +; PC64-NEXT: bl log10f +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log10 = call <1 x float> @llvm.experimental.constrained.log10.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %log10 +} + +define <2 x double> @constrained_vector_log10_v2f64() { +; PC64LE-LABEL: constrained_vector_log10_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI66_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI66_0@toc@l(3) +; PC64LE-NEXT: bl log10 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI66_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI66_1@toc@l(3) +; PC64LE-NEXT: bl log10 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 0, 1 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log10_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI66_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI66_0@toc@l(3) +; PC64LE9-NEXT: bl log10 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI66_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI66_1@toc@l(3) +; PC64LE9-NEXT: bl log10 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log10_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI66_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI66_0@toc@l(3) +; PC64-NEXT: bl log10 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI66_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI66_1@toc@l(3) +; PC64-NEXT: bl log10 +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log10 = call <2 x double> @llvm.experimental.constrained.log10.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %log10 +} + +define <3 x float> @constrained_vector_log10_v3f32() { +; PC64LE-LABEL: constrained_vector_log10_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 48 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -48(1) +; PC64LE-NEXT: addis 3, 2, .LCPI67_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI67_0@toc@l(3) +; PC64LE-NEXT: bl log10f +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI67_1@toc@ha +; PC64LE-NEXT: fmr 31, 1 +; PC64LE-NEXT: lfs 1, .LCPI67_1@toc@l(3) +; PC64LE-NEXT: bl log10f +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI67_2@toc@ha +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI67_2@toc@l(3) +; PC64LE-NEXT: bl log10f +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: addis 3, 2, .LCPI67_3@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI67_3@toc@l +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 31 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: addi 1, 1, 48 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log10_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI67_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI67_0@toc@l(3) +; PC64LE9-NEXT: bl log10f +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI67_1@toc@ha +; PC64LE9-NEXT: fmr 31, 1 +; PC64LE9-NEXT: lfs 1, .LCPI67_1@toc@l(3) +; PC64LE9-NEXT: bl log10f +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI67_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI67_2@toc@l(3) +; PC64LE9-NEXT: bl log10f +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 31 +; PC64LE9-NEXT: addis 3, 2, .LCPI67_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI67_3@toc@l +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log10_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI67_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI67_0@toc@l(3) +; PC64-NEXT: bl log10f +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI67_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI67_1@toc@l(3) +; PC64-NEXT: bl log10f +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI67_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI67_2@toc@l(3) +; PC64-NEXT: bl log10f +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log10 = call <3 x float> @llvm.experimental.constrained.log10.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %log10 +} + +define <3 x double> @constrained_vector_log10_v3f64() { +; PC64LE-LABEL: constrained_vector_log10_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI68_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI68_0@toc@l(3) +; PC64LE-NEXT: bl log10 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI68_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI68_1@toc@l(3) +; PC64LE-NEXT: bl log10 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI68_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 0, 1 +; PC64LE-NEXT: lfd 1, .LCPI68_2@toc@l(3) +; PC64LE-NEXT: bl log10 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log10_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI68_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI68_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl log10 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI68_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI68_1@toc@l(3) +; PC64LE9-NEXT: bl log10 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI68_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 0, 1 +; PC64LE9-NEXT: lfd 1, .LCPI68_2@toc@l(3) +; PC64LE9-NEXT: bl log10 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log10_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI68_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI68_0@toc@l(3) +; PC64-NEXT: bl log10 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI68_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI68_1@toc@l(3) +; PC64-NEXT: bl log10 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI68_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI68_2@toc@l(3) +; PC64-NEXT: bl log10 +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log10 = call <3 x double> @llvm.experimental.constrained.log10.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %log10 +} + +define <4 x double> @constrained_vector_log10_v4f64() { +; PC64LE-LABEL: constrained_vector_log10_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI69_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI69_0@toc@l(3) +; PC64LE-NEXT: bl log10 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI69_1@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI69_1@toc@l(3) +; PC64LE-NEXT: bl log10 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI69_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfd 1, .LCPI69_2@toc@l(3) +; PC64LE-NEXT: bl log10 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI69_3@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI69_3@toc@l(3) +; PC64LE-NEXT: bl log10 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 3, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 0, 1 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log10_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI69_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI69_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl log10 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI69_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI69_1@toc@l(3) +; PC64LE9-NEXT: bl log10 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI69_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfd 1, .LCPI69_2@toc@l(3) +; PC64LE9-NEXT: bl log10 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI69_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI69_3@toc@l(3) +; PC64LE9-NEXT: bl log10 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 3, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log10_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI69_0@toc@ha +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI69_0@toc@l(3) +; PC64-NEXT: bl log10 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI69_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI69_1@toc@l(3) +; PC64-NEXT: bl log10 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI69_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI69_2@toc@l(3) +; PC64-NEXT: bl log10 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI69_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI69_3@toc@l(3) +; PC64-NEXT: bl log10 +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log10 = call <4 x double> @llvm.experimental.constrained.log10.v4f64( + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %log10 +} + +define <1 x float> @constrained_vector_log2_v1f32() { +; PC64LE-LABEL: constrained_vector_log2_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI70_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI70_0@toc@l(3) +; PC64LE-NEXT: bl log2f +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log2_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI70_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI70_0@toc@l(3) +; PC64LE9-NEXT: bl log2f +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log2_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI70_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI70_0@toc@l(3) +; PC64-NEXT: bl log2f +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log2 = call <1 x float> @llvm.experimental.constrained.log2.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %log2 +} + +define <2 x double> @constrained_vector_log2_v2f64() { +; PC64LE-LABEL: constrained_vector_log2_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI71_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI71_0@toc@l(3) +; PC64LE-NEXT: bl log2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI71_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI71_1@toc@l(3) +; PC64LE-NEXT: bl log2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 0, 1 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log2_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI71_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI71_0@toc@l(3) +; PC64LE9-NEXT: bl log2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI71_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI71_1@toc@l(3) +; PC64LE9-NEXT: bl log2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log2_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI71_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI71_0@toc@l(3) +; PC64-NEXT: bl log2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI71_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI71_1@toc@l(3) +; PC64-NEXT: bl log2 +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log2 = call <2 x double> @llvm.experimental.constrained.log2.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %log2 +} + +define <3 x float> @constrained_vector_log2_v3f32() { +; PC64LE-LABEL: constrained_vector_log2_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 48 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -48(1) +; PC64LE-NEXT: addis 3, 2, .LCPI72_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI72_0@toc@l(3) +; PC64LE-NEXT: bl log2f +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI72_1@toc@ha +; PC64LE-NEXT: fmr 31, 1 +; PC64LE-NEXT: lfs 1, .LCPI72_1@toc@l(3) +; PC64LE-NEXT: bl log2f +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI72_2@toc@ha +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI72_2@toc@l(3) +; PC64LE-NEXT: bl log2f +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: addis 3, 2, .LCPI72_3@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI72_3@toc@l +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 31 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: addi 1, 1, 48 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log2_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI72_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI72_0@toc@l(3) +; PC64LE9-NEXT: bl log2f +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI72_1@toc@ha +; PC64LE9-NEXT: fmr 31, 1 +; PC64LE9-NEXT: lfs 1, .LCPI72_1@toc@l(3) +; PC64LE9-NEXT: bl log2f +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI72_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI72_2@toc@l(3) +; PC64LE9-NEXT: bl log2f +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 31 +; PC64LE9-NEXT: addis 3, 2, .LCPI72_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI72_3@toc@l +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log2_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI72_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI72_0@toc@l(3) +; PC64-NEXT: bl log2f +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI72_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI72_1@toc@l(3) +; PC64-NEXT: bl log2f +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI72_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI72_2@toc@l(3) +; PC64-NEXT: bl log2f +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log2 = call <3 x float> @llvm.experimental.constrained.log2.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %log2 +} + +define <3 x double> @constrained_vector_log2_v3f64() { +; PC64LE-LABEL: constrained_vector_log2_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI73_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI73_0@toc@l(3) +; PC64LE-NEXT: bl log2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI73_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI73_1@toc@l(3) +; PC64LE-NEXT: bl log2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI73_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 0, 1 +; PC64LE-NEXT: lfd 1, .LCPI73_2@toc@l(3) +; PC64LE-NEXT: bl log2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log2_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI73_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI73_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl log2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI73_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI73_1@toc@l(3) +; PC64LE9-NEXT: bl log2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI73_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 0, 1 +; PC64LE9-NEXT: lfd 1, .LCPI73_2@toc@l(3) +; PC64LE9-NEXT: bl log2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log2_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI73_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI73_0@toc@l(3) +; PC64-NEXT: bl log2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI73_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI73_1@toc@l(3) +; PC64-NEXT: bl log2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI73_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI73_2@toc@l(3) +; PC64-NEXT: bl log2 +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log2 = call <3 x double> @llvm.experimental.constrained.log2.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %log2 +} + +define <4 x double> @constrained_vector_log2_v4f64() { +; PC64LE-LABEL: constrained_vector_log2_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI74_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI74_0@toc@l(3) +; PC64LE-NEXT: bl log2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI74_1@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI74_1@toc@l(3) +; PC64LE-NEXT: bl log2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI74_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfd 1, .LCPI74_2@toc@l(3) +; PC64LE-NEXT: bl log2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI74_3@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI74_3@toc@l(3) +; PC64LE-NEXT: bl log2 +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 3, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 0, 1 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_log2_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI74_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI74_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl log2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI74_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI74_1@toc@l(3) +; PC64LE9-NEXT: bl log2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI74_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfd 1, .LCPI74_2@toc@l(3) +; PC64LE9-NEXT: bl log2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI74_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI74_3@toc@l(3) +; PC64LE9-NEXT: bl log2 +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 3, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_log2_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI74_0@toc@ha +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI74_0@toc@l(3) +; PC64-NEXT: bl log2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI74_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI74_1@toc@l(3) +; PC64-NEXT: bl log2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI74_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI74_2@toc@l(3) +; PC64-NEXT: bl log2 +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI74_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI74_3@toc@l(3) +; PC64-NEXT: bl log2 +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %log2 = call <4 x double> @llvm.experimental.constrained.log2.v4f64( + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %log2 +} + +define <1 x float> @constrained_vector_rint_v1f32() { +; PC64LE-LABEL: constrained_vector_rint_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI75_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI75_0@toc@l(3) +; PC64LE-NEXT: bl rintf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_rint_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI75_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI75_0@toc@l(3) +; PC64LE9-NEXT: bl rintf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_rint_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI75_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI75_0@toc@l(3) +; PC64-NEXT: bl rintf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %rint = call <1 x float> @llvm.experimental.constrained.rint.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %rint +} + +define <2 x double> @constrained_vector_rint_v2f64() { +; PC64LE-LABEL: constrained_vector_rint_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI76_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI76_0@toc@l(3) +; PC64LE-NEXT: bl rint +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI76_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI76_1@toc@l(3) +; PC64LE-NEXT: bl rint +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 1, 0 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_rint_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI76_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI76_0@toc@l(3) +; PC64LE9-NEXT: bl rint +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI76_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI76_1@toc@l(3) +; PC64LE9-NEXT: bl rint +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_rint_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI76_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI76_0@toc@l(3) +; PC64-NEXT: bl rint +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI76_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI76_1@toc@l(3) +; PC64-NEXT: bl rint +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %rint = call <2 x double> @llvm.experimental.constrained.rint.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %rint +} + +define <3 x float> @constrained_vector_rint_v3f32() { +; PC64LE-LABEL: constrained_vector_rint_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 48 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -48(1) +; PC64LE-NEXT: addis 3, 2, .LCPI77_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI77_0@toc@l(3) +; PC64LE-NEXT: bl rintf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI77_1@toc@ha +; PC64LE-NEXT: fmr 31, 1 +; PC64LE-NEXT: lfs 1, .LCPI77_1@toc@l(3) +; PC64LE-NEXT: bl rintf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI77_2@toc@ha +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI77_2@toc@l(3) +; PC64LE-NEXT: bl rintf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: addis 3, 2, .LCPI77_3@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI77_3@toc@l +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 31 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: addi 1, 1, 48 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_rint_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI77_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI77_0@toc@l(3) +; PC64LE9-NEXT: bl rintf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI77_1@toc@ha +; PC64LE9-NEXT: fmr 31, 1 +; PC64LE9-NEXT: lfs 1, .LCPI77_1@toc@l(3) +; PC64LE9-NEXT: bl rintf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI77_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI77_2@toc@l(3) +; PC64LE9-NEXT: bl rintf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 31 +; PC64LE9-NEXT: addis 3, 2, .LCPI77_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI77_3@toc@l +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_rint_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI77_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI77_0@toc@l(3) +; PC64-NEXT: bl rintf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI77_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI77_1@toc@l(3) +; PC64-NEXT: bl rintf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI77_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI77_2@toc@l(3) +; PC64-NEXT: bl rintf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr + entry: + %rint = call <3 x float> @llvm.experimental.constrained.rint.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %rint +} + +define <3 x double> @constrained_vector_rint_v3f64() { +; PC64LE-LABEL: constrained_vector_rint_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI78_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI78_0@toc@l(3) +; PC64LE-NEXT: bl rint +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI78_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI78_1@toc@l(3) +; PC64LE-NEXT: bl rint +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI78_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 0, 1 +; PC64LE-NEXT: lfd 1, .LCPI78_2@toc@l(3) +; PC64LE-NEXT: bl rint +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_rint_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI78_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI78_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl rint +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI78_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI78_1@toc@l(3) +; PC64LE9-NEXT: bl rint +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI78_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 0, 1 +; PC64LE9-NEXT: lfd 1, .LCPI78_2@toc@l(3) +; PC64LE9-NEXT: bl rint +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_rint_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI78_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI78_0@toc@l(3) +; PC64-NEXT: bl rint +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI78_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI78_1@toc@l(3) +; PC64-NEXT: bl rint +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI78_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI78_2@toc@l(3) +; PC64-NEXT: bl rint +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %rint = call <3 x double> @llvm.experimental.constrained.rint.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %rint +} + +define <4 x double> @constrained_vector_rint_v4f64() { +; PC64LE-LABEL: constrained_vector_rint_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI79_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI79_0@toc@l(3) +; PC64LE-NEXT: bl rint +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI79_1@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI79_1@toc@l(3) +; PC64LE-NEXT: bl rint +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI79_2@toc@ha +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfd 1, .LCPI79_2@toc@l(3) +; PC64LE-NEXT: bl rint +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI79_3@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI79_3@toc@l(3) +; PC64LE-NEXT: bl rint +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 2, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 35, 1, 0 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_rint_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI79_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI79_0@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl rint +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI79_1@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI79_1@toc@l(3) +; PC64LE9-NEXT: bl rint +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI79_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfd 1, .LCPI79_2@toc@l(3) +; PC64LE9-NEXT: bl rint +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI79_3@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfd 1, .LCPI79_3@toc@l(3) +; PC64LE9-NEXT: bl rint +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 2, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 35, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_rint_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI79_0@toc@ha +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI79_0@toc@l(3) +; PC64-NEXT: bl rint +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI79_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI79_1@toc@l(3) +; PC64-NEXT: bl rint +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI79_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI79_2@toc@l(3) +; PC64-NEXT: bl rint +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI79_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI79_3@toc@l(3) +; PC64-NEXT: bl rint +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %rint = call <4 x double> @llvm.experimental.constrained.rint.v4f64( + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %rint +} + +define <1 x float> @constrained_vector_nearbyint_v1f32() { +; PC64LE-LABEL: constrained_vector_nearbyint_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI80_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI80_0@toc@l(3) +; PC64LE-NEXT: bl nearbyintf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_nearbyint_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI80_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI80_0@toc@l(3) +; PC64LE9-NEXT: bl nearbyintf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_nearbyint_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI80_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI80_0@toc@l(3) +; PC64-NEXT: bl nearbyintf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %nearby = call <1 x float> @llvm.experimental.constrained.nearbyint.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %nearby +} + +define <2 x double> @constrained_vector_nearbyint_v2f64() { +; PC64LE-LABEL: constrained_vector_nearbyint_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI81_0@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI81_0@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xvrdpic 34, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_nearbyint_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI81_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI81_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvrdpic 34, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_nearbyint_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI81_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI81_0@toc@l(3) +; PC64-NEXT: bl nearbyint +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI81_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI81_1@toc@l(3) +; PC64-NEXT: bl nearbyint +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %nearby = call <2 x double> @llvm.experimental.constrained.nearbyint.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %nearby +} + +define <3 x float> @constrained_vector_nearbyint_v3f32() { +; PC64LE-LABEL: constrained_vector_nearbyint_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 48 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -48(1) +; PC64LE-NEXT: addis 3, 2, .LCPI82_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI82_0@toc@l(3) +; PC64LE-NEXT: bl nearbyintf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI82_1@toc@ha +; PC64LE-NEXT: fmr 31, 1 +; PC64LE-NEXT: lfs 1, .LCPI82_1@toc@l(3) +; PC64LE-NEXT: bl nearbyintf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI82_2@toc@ha +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI82_2@toc@l(3) +; PC64LE-NEXT: bl nearbyintf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: addis 3, 2, .LCPI82_3@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI82_3@toc@l +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 31 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: addi 1, 1, 48 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_nearbyint_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI82_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI82_0@toc@l(3) +; PC64LE9-NEXT: bl nearbyintf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI82_1@toc@ha +; PC64LE9-NEXT: fmr 31, 1 +; PC64LE9-NEXT: lfs 1, .LCPI82_1@toc@l(3) +; PC64LE9-NEXT: bl nearbyintf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI82_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI82_2@toc@l(3) +; PC64LE9-NEXT: bl nearbyintf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 31 +; PC64LE9-NEXT: addis 3, 2, .LCPI82_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI82_3@toc@l +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_nearbyint_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI82_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI82_0@toc@l(3) +; PC64-NEXT: bl nearbyintf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI82_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI82_1@toc@l(3) +; PC64-NEXT: bl nearbyintf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI82_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI82_2@toc@l(3) +; PC64-NEXT: bl nearbyintf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %nearby = call <3 x float> @llvm.experimental.constrained.nearbyint.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %nearby +} + +define <3 x double> @constrained_vector_nearby_v3f64() { +; PC64LE-LABEL: constrained_vector_nearby_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI83_0@toc@ha +; PC64LE-NEXT: lfd 1, .LCPI83_0@toc@l(3) +; PC64LE-NEXT: bl nearbyint +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI83_1@toc@ha +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI83_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xvrdpic 2, 0 +; PC64LE-NEXT: xxswapd 0, 2 +; PC64LE-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE-NEXT: fmr 1, 0 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_nearby_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI83_0@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI83_0@toc@l(3) +; PC64LE9-NEXT: bl nearbyint +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI83_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI83_1@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvrdpic 2, 0 +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xxswapd 1, 2 +; PC64LE9-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE9-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_nearby_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI83_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI83_0@toc@l(3) +; PC64-NEXT: bl nearbyint +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI83_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI83_1@toc@l(3) +; PC64-NEXT: bl nearbyint +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI83_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI83_2@toc@l(3) +; PC64-NEXT: bl nearbyint +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %nearby = call <3 x double> @llvm.experimental.constrained.nearbyint.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %nearby +} + +define <4 x double> @constrained_vector_nearbyint_v4f64() { +; PC64LE-LABEL: constrained_vector_nearbyint_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI84_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI84_1@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI84_0@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: addi 3, 4, .LCPI84_1@toc@l +; PC64LE-NEXT: lxvd2x 1, 0, 3 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xxswapd 1, 1 +; PC64LE-NEXT: xvrdpic 34, 0 +; PC64LE-NEXT: xvrdpic 35, 1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_nearbyint_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI84_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI84_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI84_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI84_1@toc@l +; PC64LE9-NEXT: xvrdpic 34, 0 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvrdpic 35, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_nearbyint_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI84_0@toc@ha +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI84_0@toc@l(3) +; PC64-NEXT: bl nearbyint +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI84_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI84_1@toc@l(3) +; PC64-NEXT: bl nearbyint +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI84_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfd 1, .LCPI84_2@toc@l(3) +; PC64-NEXT: bl nearbyint +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI84_3@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfd 1, .LCPI84_3@toc@l(3) +; PC64-NEXT: bl nearbyint +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %nearby = call <4 x double> @llvm.experimental.constrained.nearbyint.v4f64( + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %nearby +} + +define <1 x float> @constrained_vector_maxnum_v1f32() { +; PC64LE-LABEL: constrained_vector_maxnum_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI85_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI85_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI85_0@toc@l(3) +; PC64LE-NEXT: lfs 2, .LCPI85_1@toc@l(4) +; PC64LE-NEXT: bl fmaxf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_maxnum_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI85_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI85_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI85_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI85_1@toc@l(3) +; PC64LE9-NEXT: bl fmaxf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_maxnum_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI85_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI85_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI85_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI85_1@toc@l(3) +; PC64-NEXT: bl fmaxf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %max = call <1 x float> @llvm.experimental.constrained.maxnum.v1f32( + <1 x float> , <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %max +} + +define <2 x double> @constrained_vector_maxnum_v2f64() { +; PC64LE-LABEL: constrained_vector_maxnum_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI86_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI86_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI86_0@toc@l(3) +; PC64LE-NEXT: lfs 2, .LCPI86_1@toc@l(4) +; PC64LE-NEXT: bl fmax +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: addis 4, 2, .LCPI86_3@toc@ha +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI86_2@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI86_3@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI86_2@toc@l(3) +; PC64LE-NEXT: bl fmax +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 1, 0 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_maxnum_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI86_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI86_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI86_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI86_1@toc@l(3) +; PC64LE9-NEXT: bl fmax +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI86_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI86_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI86_3@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI86_3@toc@l(3) +; PC64LE9-NEXT: bl fmax +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_maxnum_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI86_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI86_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI86_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI86_1@toc@l(3) +; PC64-NEXT: bl fmax +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI86_2@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI86_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI86_3@toc@ha +; PC64-NEXT: lfs 2, .LCPI86_3@toc@l(3) +; PC64-NEXT: bl fmax +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %max = call <2 x double> @llvm.experimental.constrained.maxnum.v2f64( + <2 x double> , + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %max +} + +define <3 x float> @constrained_vector_maxnum_v3f32() { +; PC64LE-LABEL: constrained_vector_maxnum_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f29, -24 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 29, -24(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: addis 4, 2, .LCPI87_1@toc@ha +; PC64LE-NEXT: addis 3, 2, .LCPI87_0@toc@ha +; PC64LE-NEXT: lfs 31, .LCPI87_1@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI87_0@toc@l(3) +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: bl fmaxf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI87_2@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI87_3@toc@ha +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI87_2@toc@l(3) +; PC64LE-NEXT: lfs 2, .LCPI87_3@toc@l(4) +; PC64LE-NEXT: bl fmaxf +; PC64LE-NEXT: nop +; PC64LE-NEXT: fmr 29, 1 +; PC64LE-NEXT: addis 3, 2, .LCPI87_4@toc@ha +; PC64LE-NEXT: fmr 1, 31 +; PC64LE-NEXT: lfs 2, .LCPI87_4@toc@l(3) +; PC64LE-NEXT: bl fmaxf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 29 +; PC64LE-NEXT: addis 3, 2, .LCPI87_5@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI87_5@toc@l +; PC64LE-NEXT: lvx 4, 0, 3 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 3, 2, 4 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 29, -24(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_maxnum_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f29, -24 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 29, -24(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI87_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI87_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI87_1@toc@ha +; PC64LE9-NEXT: lfs 31, .LCPI87_1@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl fmaxf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI87_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI87_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI87_3@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI87_3@toc@l(3) +; PC64LE9-NEXT: bl fmaxf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI87_4@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI87_4@toc@l(3) +; PC64LE9-NEXT: fmr 29, 1 +; PC64LE9-NEXT: fmr 1, 31 +; PC64LE9-NEXT: bl fmaxf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 29 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: addis 3, 2, .LCPI87_5@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI87_5@toc@l +; PC64LE9-NEXT: lxvx 36, 0, 3 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 3, 2, 4 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 29, -24(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_maxnum_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI87_0@toc@ha +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 31, .LCPI87_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI87_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI87_1@toc@l(3) +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: bl fmaxf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI87_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI87_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI87_3@toc@ha +; PC64-NEXT: lfs 2, .LCPI87_3@toc@l(3) +; PC64-NEXT: bl fmaxf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI87_4@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfs 1, .LCPI87_4@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fmaxf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 30 +; PC64-NEXT: fmr 2, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %max = call <3 x float> @llvm.experimental.constrained.maxnum.v3f32( + <3 x float> , + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %max +} + +define <3 x double> @constrained_vector_max_v3f64() { +; PC64LE-LABEL: constrained_vector_max_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: addis 4, 2, .LCPI88_1@toc@ha +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI88_0@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI88_1@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI88_0@toc@l(3) +; PC64LE-NEXT: bl fmax +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: addis 4, 2, .LCPI88_3@toc@ha +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI88_2@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI88_3@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI88_2@toc@l(3) +; PC64LE-NEXT: bl fmax +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: addis 4, 2, .LCPI88_5@toc@ha +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI88_4@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI88_5@toc@l(4) +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfs 1, .LCPI88_4@toc@l(3) +; PC64LE-NEXT: bl fmax +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_max_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI88_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI88_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI88_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI88_1@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl fmax +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI88_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI88_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI88_3@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI88_3@toc@l(3) +; PC64LE9-NEXT: bl fmax +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI88_4@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfs 1, .LCPI88_4@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI88_5@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI88_5@toc@l(3) +; PC64LE9-NEXT: bl fmax +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_max_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI88_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI88_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI88_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI88_1@toc@l(3) +; PC64-NEXT: bl fmax +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI88_2@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI88_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI88_3@toc@ha +; PC64-NEXT: lfs 2, .LCPI88_3@toc@l(3) +; PC64-NEXT: bl fmax +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI88_4@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI88_4@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI88_5@toc@ha +; PC64-NEXT: lfs 2, .LCPI88_5@toc@l(3) +; PC64-NEXT: bl fmax +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %max = call <3 x double> @llvm.experimental.constrained.maxnum.v3f64( + <3 x double> , + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %max +} + +define <4 x double> @constrained_vector_maxnum_v4f64() { +; PC64LE-LABEL: constrained_vector_maxnum_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: addis 4, 2, .LCPI89_1@toc@ha +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI89_0@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI89_1@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI89_0@toc@l(3) +; PC64LE-NEXT: bl fmax +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: addis 4, 2, .LCPI89_3@toc@ha +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI89_2@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI89_3@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI89_2@toc@l(3) +; PC64LE-NEXT: bl fmax +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: addis 4, 2, .LCPI89_5@toc@ha +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI89_4@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI89_5@toc@l(4) +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfs 1, .LCPI89_4@toc@l(3) +; PC64LE-NEXT: bl fmax +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: addis 4, 2, .LCPI89_7@toc@ha +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI89_6@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI89_7@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI89_6@toc@l(3) +; PC64LE-NEXT: bl fmax +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 2, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 35, 1, 0 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_maxnum_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI89_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI89_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI89_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI89_1@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl fmax +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI89_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI89_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI89_3@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI89_3@toc@l(3) +; PC64LE9-NEXT: bl fmax +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI89_4@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfs 1, .LCPI89_4@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI89_5@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI89_5@toc@l(3) +; PC64LE9-NEXT: bl fmax +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI89_6@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI89_6@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI89_7@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI89_7@toc@l(3) +; PC64LE9-NEXT: bl fmax +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 2, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 35, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_maxnum_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI89_0@toc@ha +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI89_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI89_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI89_1@toc@l(3) +; PC64-NEXT: bl fmax +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI89_2@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI89_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI89_3@toc@ha +; PC64-NEXT: lfs 2, .LCPI89_3@toc@l(3) +; PC64-NEXT: bl fmax +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI89_4@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI89_4@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI89_5@toc@ha +; PC64-NEXT: lfs 2, .LCPI89_5@toc@l(3) +; PC64-NEXT: bl fmax +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI89_6@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfs 1, .LCPI89_6@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI89_7@toc@ha +; PC64-NEXT: lfs 2, .LCPI89_7@toc@l(3) +; PC64-NEXT: bl fmax +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %max = call <4 x double> @llvm.experimental.constrained.maxnum.v4f64( + <4 x double> , + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %max +} + +define <1 x float> @constrained_vector_minnum_v1f32() { +; PC64LE-LABEL: constrained_vector_minnum_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -32(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 32 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI90_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI90_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI90_0@toc@l(3) +; PC64LE-NEXT: lfs 2, .LCPI90_1@toc@l(4) +; PC64LE-NEXT: bl fminf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: addi 1, 1, 32 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_minnum_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -32(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 32 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI90_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI90_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI90_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI90_1@toc@l(3) +; PC64LE9-NEXT: bl fminf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: addi 1, 1, 32 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_minnum_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI90_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI90_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI90_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI90_1@toc@l(3) +; PC64-NEXT: bl fminf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr + entry: + %min = call <1 x float> @llvm.experimental.constrained.minnum.v1f32( + <1 x float> , <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %min +} + +define <2 x double> @constrained_vector_minnum_v2f64() { +; PC64LE-LABEL: constrained_vector_minnum_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: addis 3, 2, .LCPI91_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI91_1@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI91_0@toc@l(3) +; PC64LE-NEXT: lfs 2, .LCPI91_1@toc@l(4) +; PC64LE-NEXT: bl fmin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: addis 4, 2, .LCPI91_3@toc@ha +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI91_2@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI91_3@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI91_2@toc@l(3) +; PC64LE-NEXT: bl fmin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 34, 1, 0 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_minnum_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -48(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 48 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: addis 3, 2, .LCPI91_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI91_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI91_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI91_1@toc@l(3) +; PC64LE9-NEXT: bl fmin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI91_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI91_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI91_3@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI91_3@toc@l(3) +; PC64LE9-NEXT: bl fmin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 34, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 48 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_minnum_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI91_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI91_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI91_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI91_1@toc@l(3) +; PC64-NEXT: bl fmin +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI91_2@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI91_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI91_3@toc@ha +; PC64-NEXT: lfs 2, .LCPI91_3@toc@l(3) +; PC64-NEXT: bl fmin +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %min = call <2 x double> @llvm.experimental.constrained.minnum.v2f64( + <2 x double> , + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %min +} + +define <3 x float> @constrained_vector_minnum_v3f32() { +; PC64LE-LABEL: constrained_vector_minnum_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: .cfi_def_cfa_offset 64 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset f29, -24 +; PC64LE-NEXT: .cfi_offset f30, -16 +; PC64LE-NEXT: .cfi_offset f31, -8 +; PC64LE-NEXT: stfd 29, -24(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -64(1) +; PC64LE-NEXT: addis 4, 2, .LCPI92_1@toc@ha +; PC64LE-NEXT: addis 3, 2, .LCPI92_0@toc@ha +; PC64LE-NEXT: lfs 31, .LCPI92_1@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI92_0@toc@l(3) +; PC64LE-NEXT: fmr 2, 31 +; PC64LE-NEXT: bl fminf +; PC64LE-NEXT: nop +; PC64LE-NEXT: addis 3, 2, .LCPI92_2@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI92_3@toc@ha +; PC64LE-NEXT: fmr 30, 1 +; PC64LE-NEXT: lfs 1, .LCPI92_2@toc@l(3) +; PC64LE-NEXT: lfs 2, .LCPI92_3@toc@l(4) +; PC64LE-NEXT: bl fminf +; PC64LE-NEXT: nop +; PC64LE-NEXT: fmr 29, 1 +; PC64LE-NEXT: addis 3, 2, .LCPI92_4@toc@ha +; PC64LE-NEXT: fmr 1, 31 +; PC64LE-NEXT: lfs 2, .LCPI92_4@toc@l(3) +; PC64LE-NEXT: bl fminf +; PC64LE-NEXT: nop +; PC64LE-NEXT: xscvdpspn 0, 29 +; PC64LE-NEXT: addis 3, 2, .LCPI92_5@toc@ha +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI92_5@toc@l +; PC64LE-NEXT: lvx 4, 0, 3 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 30 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 2, 3 +; PC64LE-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 3, 2, 4 +; PC64LE-NEXT: addi 1, 1, 64 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE-NEXT: lfd 29, -24(1) # 8-byte Folded Reload +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_minnum_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset f29, -24 +; PC64LE9-NEXT: .cfi_offset f30, -16 +; PC64LE9-NEXT: .cfi_offset f31, -8 +; PC64LE9-NEXT: stfd 29, -24(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 30, -16(1) # 8-byte Folded Spill +; PC64LE9-NEXT: stfd 31, -8(1) # 8-byte Folded Spill +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: addis 3, 2, .LCPI92_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI92_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI92_1@toc@ha +; PC64LE9-NEXT: lfs 31, .LCPI92_1@toc@l(3) +; PC64LE9-NEXT: fmr 2, 31 +; PC64LE9-NEXT: bl fminf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI92_2@toc@ha +; PC64LE9-NEXT: fmr 30, 1 +; PC64LE9-NEXT: lfs 1, .LCPI92_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI92_3@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI92_3@toc@l(3) +; PC64LE9-NEXT: bl fminf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI92_4@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI92_4@toc@l(3) +; PC64LE9-NEXT: fmr 29, 1 +; PC64LE9-NEXT: fmr 1, 31 +; PC64LE9-NEXT: bl fminf +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: xscvdpspn 0, 1 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 29 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: xscvdpspn 0, 30 +; PC64LE9-NEXT: addis 3, 2, .LCPI92_5@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI92_5@toc@l +; PC64LE9-NEXT: lxvx 36, 0, 3 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 3, 2, 4 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: lfd 31, -8(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 30, -16(1) # 8-byte Folded Reload +; PC64LE9-NEXT: lfd 29, -24(1) # 8-byte Folded Reload +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_minnum_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI92_0@toc@ha +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 31, .LCPI92_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI92_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI92_1@toc@l(3) +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: bl fminf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI92_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI92_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI92_3@toc@ha +; PC64-NEXT: lfs 2, .LCPI92_3@toc@l(3) +; PC64-NEXT: bl fminf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI92_4@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfs 1, .LCPI92_4@toc@l(3) +; PC64-NEXT: fmr 2, 31 +; PC64-NEXT: bl fminf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 30 +; PC64-NEXT: fmr 2, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %min = call <3 x float> @llvm.experimental.constrained.minnum.v3f32( + <3 x float> , + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %min +} + +define <3 x double> @constrained_vector_min_v3f64() { +; PC64LE-LABEL: constrained_vector_min_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: addis 4, 2, .LCPI93_1@toc@ha +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI93_0@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI93_1@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI93_0@toc@l(3) +; PC64LE-NEXT: bl fmin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: addis 4, 2, .LCPI93_3@toc@ha +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI93_2@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI93_3@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI93_2@toc@l(3) +; PC64LE-NEXT: bl fmin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: addis 4, 2, .LCPI93_5@toc@ha +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI93_4@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI93_5@toc@l(4) +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfs 1, .LCPI93_4@toc@l(3) +; PC64LE-NEXT: bl fmin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: fmr 3, 1 +; PC64LE-NEXT: xxlor 1, 63, 63 +; PC64LE-NEXT: xxlor 2, 63, 63 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_min_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI93_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI93_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI93_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI93_1@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl fmin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI93_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI93_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI93_3@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI93_3@toc@l(3) +; PC64LE9-NEXT: bl fmin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI93_4@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfs 1, .LCPI93_4@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI93_5@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI93_5@toc@l(3) +; PC64LE9-NEXT: bl fmin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: fmr 3, 1 +; PC64LE9-NEXT: xscpsgndp 1, 63, 63 +; PC64LE9-NEXT: xscpsgndp 2, 63, 63 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_min_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI93_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI93_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI93_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI93_1@toc@l(3) +; PC64-NEXT: bl fmin +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI93_2@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI93_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI93_3@toc@ha +; PC64-NEXT: lfs 2, .LCPI93_3@toc@l(3) +; PC64-NEXT: bl fmin +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI93_4@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI93_4@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI93_5@toc@ha +; PC64-NEXT: lfs 2, .LCPI93_5@toc@l(3) +; PC64-NEXT: bl fmin +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %min = call <3 x double> @llvm.experimental.constrained.minnum.v3f64( + <3 x double> , + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %min +} + +define <4 x double> @constrained_vector_minnum_v4f64() { +; PC64LE-LABEL: constrained_vector_minnum_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: mflr 0 +; PC64LE-NEXT: std 0, 16(1) +; PC64LE-NEXT: stdu 1, -80(1) +; PC64LE-NEXT: .cfi_def_cfa_offset 80 +; PC64LE-NEXT: .cfi_offset lr, 16 +; PC64LE-NEXT: .cfi_offset v31, -16 +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: addis 4, 2, .LCPI94_1@toc@ha +; PC64LE-NEXT: stxvd2x 63, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI94_0@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI94_1@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI94_0@toc@l(3) +; PC64LE-NEXT: bl fmin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: addis 4, 2, .LCPI94_3@toc@ha +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI94_2@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI94_3@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI94_2@toc@l(3) +; PC64LE-NEXT: bl fmin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: addis 4, 2, .LCPI94_5@toc@ha +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: addis 3, 2, .LCPI94_4@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI94_5@toc@l(4) +; PC64LE-NEXT: xxmrghd 63, 1, 0 +; PC64LE-NEXT: lfs 1, .LCPI94_4@toc@l(3) +; PC64LE-NEXT: bl fmin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: addis 4, 2, .LCPI94_7@toc@ha +; PC64LE-NEXT: stxvd2x 1, 1, 3 # 16-byte Folded Spill +; PC64LE-NEXT: addis 3, 2, .LCPI94_6@toc@ha +; PC64LE-NEXT: lfs 2, .LCPI94_7@toc@l(4) +; PC64LE-NEXT: lfs 1, .LCPI94_6@toc@l(3) +; PC64LE-NEXT: bl fmin +; PC64LE-NEXT: nop +; PC64LE-NEXT: li 3, 48 +; PC64LE-NEXT: vmr 2, 31 +; PC64LE-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE-NEXT: lxvd2x 0, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: li 3, 64 +; PC64LE-NEXT: lxvd2x 63, 1, 3 # 16-byte Folded Reload +; PC64LE-NEXT: xxmrghd 35, 1, 0 +; PC64LE-NEXT: addi 1, 1, 80 +; PC64LE-NEXT: ld 0, 16(1) +; PC64LE-NEXT: mtlr 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_minnum_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: mflr 0 +; PC64LE9-NEXT: std 0, 16(1) +; PC64LE9-NEXT: stdu 1, -64(1) +; PC64LE9-NEXT: .cfi_def_cfa_offset 64 +; PC64LE9-NEXT: .cfi_offset lr, 16 +; PC64LE9-NEXT: .cfi_offset v31, -16 +; PC64LE9-NEXT: addis 3, 2, .LCPI94_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI94_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI94_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI94_1@toc@l(3) +; PC64LE9-NEXT: stxv 63, 48(1) # 16-byte Folded Spill +; PC64LE9-NEXT: bl fmin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI94_2@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI94_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI94_3@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI94_3@toc@l(3) +; PC64LE9-NEXT: bl fmin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: addis 3, 2, .LCPI94_4@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 63, 1, 0 +; PC64LE9-NEXT: lfs 1, .LCPI94_4@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI94_5@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI94_5@toc@l(3) +; PC64LE9-NEXT: bl fmin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: addis 3, 2, .LCPI94_6@toc@ha +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: stxv 1, 32(1) # 16-byte Folded Spill +; PC64LE9-NEXT: lfs 1, .LCPI94_6@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI94_7@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI94_7@toc@l(3) +; PC64LE9-NEXT: bl fmin +; PC64LE9-NEXT: nop +; PC64LE9-NEXT: lxv 0, 32(1) # 16-byte Folded Reload +; PC64LE9-NEXT: vmr 2, 31 +; PC64LE9-NEXT: lxv 63, 48(1) # 16-byte Folded Reload +; PC64LE9-NEXT: # kill: def $f1 killed $f1 def $vsl1 +; PC64LE9-NEXT: xxmrghd 35, 1, 0 +; PC64LE9-NEXT: addi 1, 1, 64 +; PC64LE9-NEXT: ld 0, 16(1) +; PC64LE9-NEXT: mtlr 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_minnum_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -144(1) +; PC64-NEXT: .cfi_def_cfa_offset 144 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f29, -24 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI94_0@toc@ha +; PC64-NEXT: stfd 29, 120(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 30, 128(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 136(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI94_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI94_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI94_1@toc@l(3) +; PC64-NEXT: bl fmin +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI94_2@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI94_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI94_3@toc@ha +; PC64-NEXT: lfs 2, .LCPI94_3@toc@l(3) +; PC64-NEXT: bl fmin +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI94_4@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI94_4@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI94_5@toc@ha +; PC64-NEXT: lfs 2, .LCPI94_5@toc@l(3) +; PC64-NEXT: bl fmin +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI94_6@toc@ha +; PC64-NEXT: fmr 29, 1 +; PC64-NEXT: lfs 1, .LCPI94_6@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI94_7@toc@ha +; PC64-NEXT: lfs 2, .LCPI94_7@toc@l(3) +; PC64-NEXT: bl fmin +; PC64-NEXT: nop +; PC64-NEXT: fmr 4, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: fmr 3, 29 +; PC64-NEXT: lfd 31, 136(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 128(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 29, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 144 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %min = call <4 x double> @llvm.experimental.constrained.minnum.v4f64( + <4 x double> , + <4 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x double> %min +} + +define <1 x float> @constrained_vector_fptrunc_v1f64() { +; PC64LE-LABEL: constrained_vector_fptrunc_v1f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI95_0@toc@ha +; PC64LE-NEXT: lfd 0, .LCPI95_0@toc@l(3) +; PC64LE-NEXT: frsp 0, 0 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fptrunc_v1f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI95_0@toc@ha +; PC64LE9-NEXT: lfd 0, .LCPI95_0@toc@l(3) +; PC64LE9-NEXT: frsp 0, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fptrunc_v1f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI95_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI95_0@toc@l(3) +; PC64-NEXT: frsp 1, 0 +; PC64-NEXT: blr +entry: + %result = call <1 x float> @llvm.experimental.constrained.fptrunc.v1f32.v1f64( + <1 x double>, + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %result +} + +define <2 x float> @constrained_vector_fptrunc_v2f64() { +; PC64LE-LABEL: constrained_vector_fptrunc_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI96_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI96_1@toc@ha +; PC64LE-NEXT: lfd 0, .LCPI96_0@toc@l(3) +; PC64LE-NEXT: lfd 1, .LCPI96_1@toc@l(4) +; PC64LE-NEXT: frsp 0, 0 +; PC64LE-NEXT: frsp 1, 1 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 3, 2 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fptrunc_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI96_0@toc@ha +; PC64LE9-NEXT: lfd 0, .LCPI96_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI96_1@toc@ha +; PC64LE9-NEXT: frsp 0, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: lfd 0, .LCPI96_1@toc@l(3) +; PC64LE9-NEXT: frsp 0, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fptrunc_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI96_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI96_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI96_1@toc@ha +; PC64-NEXT: lfd 2, .LCPI96_1@toc@l(3) +; PC64-NEXT: frsp 1, 0 +; PC64-NEXT: frsp 2, 2 +; PC64-NEXT: blr +entry: + %result = call <2 x float> @llvm.experimental.constrained.fptrunc.v2f32.v2f64( + <2 x double>, + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x float> %result +} + +define <3 x float> @constrained_vector_fptrunc_v3f64() { +; PC64LE-LABEL: constrained_vector_fptrunc_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI97_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI97_1@toc@ha +; PC64LE-NEXT: lfd 0, .LCPI97_0@toc@l(3) +; PC64LE-NEXT: lfd 1, .LCPI97_1@toc@l(4) +; PC64LE-NEXT: addis 3, 2, .LCPI97_3@toc@ha +; PC64LE-NEXT: frsp 0, 0 +; PC64LE-NEXT: lfd 2, .LCPI97_3@toc@l(3) +; PC64LE-NEXT: addis 3, 2, .LCPI97_2@toc@ha +; PC64LE-NEXT: frsp 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI97_2@toc@l +; PC64LE-NEXT: frsp 2, 2 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 2 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 3, 2 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fptrunc_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI97_0@toc@ha +; PC64LE9-NEXT: lfd 0, .LCPI97_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI97_1@toc@ha +; PC64LE9-NEXT: frsp 0, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: lfd 0, .LCPI97_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI97_2@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI97_2@toc@l +; PC64LE9-NEXT: frsp 0, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 35, 0, 0, 1 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: addis 3, 2, .LCPI97_3@toc@ha +; PC64LE9-NEXT: lfd 0, .LCPI97_3@toc@l(3) +; PC64LE9-NEXT: frsp 0, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fptrunc_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI97_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI97_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI97_1@toc@ha +; PC64-NEXT: lfd 2, .LCPI97_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI97_2@toc@ha +; PC64-NEXT: lfd 3, .LCPI97_2@toc@l(3) +; PC64-NEXT: frsp 1, 0 +; PC64-NEXT: frsp 2, 2 +; PC64-NEXT: frsp 3, 3 +; PC64-NEXT: blr +entry: + %result = call <3 x float> @llvm.experimental.constrained.fptrunc.v3f32.v3f64( + <3 x double>, + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %result +} + +define <4 x float> @constrained_vector_fptrunc_v4f64() { +; PC64LE-LABEL: constrained_vector_fptrunc_v4f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI98_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI98_1@toc@ha +; PC64LE-NEXT: addis 5, 2, .LCPI98_2@toc@ha +; PC64LE-NEXT: addis 6, 2, .LCPI98_3@toc@ha +; PC64LE-NEXT: lfd 0, .LCPI98_0@toc@l(3) +; PC64LE-NEXT: lfd 1, .LCPI98_1@toc@l(4) +; PC64LE-NEXT: lfd 2, .LCPI98_2@toc@l(5) +; PC64LE-NEXT: lfd 3, .LCPI98_3@toc@l(6) +; PC64LE-NEXT: xxmrghd 0, 1, 0 +; PC64LE-NEXT: xxmrghd 1, 3, 2 +; PC64LE-NEXT: xvcvdpsp 34, 0 +; PC64LE-NEXT: xvcvdpsp 35, 1 +; PC64LE-NEXT: vmrgew 2, 3, 2 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fptrunc_v4f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI98_0@toc@ha +; PC64LE9-NEXT: lfd 0, .LCPI98_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI98_1@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI98_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI98_2@toc@ha +; PC64LE9-NEXT: xxmrghd 0, 1, 0 +; PC64LE9-NEXT: xvcvdpsp 34, 0 +; PC64LE9-NEXT: lfd 0, .LCPI98_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI98_3@toc@ha +; PC64LE9-NEXT: lfd 1, .LCPI98_3@toc@l(3) +; PC64LE9-NEXT: xxmrghd 0, 1, 0 +; PC64LE9-NEXT: xvcvdpsp 35, 0 +; PC64LE9-NEXT: vmrgew 2, 3, 2 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fptrunc_v4f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI98_0@toc@ha +; PC64-NEXT: lfd 0, .LCPI98_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI98_1@toc@ha +; PC64-NEXT: lfd 2, .LCPI98_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI98_2@toc@ha +; PC64-NEXT: lfd 3, .LCPI98_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI98_3@toc@ha +; PC64-NEXT: frsp 1, 0 +; PC64-NEXT: lfd 4, .LCPI98_3@toc@l(3) +; PC64-NEXT: frsp 2, 2 +; PC64-NEXT: frsp 3, 3 +; PC64-NEXT: frsp 4, 4 +; PC64-NEXT: blr +entry: + %result = call <4 x float> @llvm.experimental.constrained.fptrunc.v4f32.v4f64( + <4 x double>, + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <4 x float> %result +} + +define <1 x double> @constrained_vector_fpext_v1f32() { +; PC64LE-LABEL: constrained_vector_fpext_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI99_0@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI99_0@toc@l(3) +; PC64LE-NEXT: xxspltd 34, 0, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fpext_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI99_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI99_0@toc@l(3) +; PC64LE9-NEXT: xxspltd 34, 0, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fpext_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI99_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI99_0@toc@l(3) +; PC64-NEXT: blr +entry: + %result = call <1 x double> @llvm.experimental.constrained.fpext.v1f64.v1f32( + <1 x float>, + metadata !"fpexcept.strict") + ret <1 x double> %result +} + +define <2 x double> @constrained_vector_fpext_v2f32() { +; PC64LE-LABEL: constrained_vector_fpext_v2f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI100_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI100_1@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI100_0@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI100_1@toc@l(4) +; PC64LE-NEXT: xxmrghd 34, 1, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fpext_v2f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI100_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI100_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI100_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI100_1@toc@l(3) +; PC64LE9-NEXT: xxmrghd 34, 1, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fpext_v2f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI100_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI100_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI100_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI100_1@toc@l(3) +; PC64-NEXT: blr +entry: + %result = call <2 x double> @llvm.experimental.constrained.fpext.v2f64.v2f32( + <2 x float>, + metadata !"fpexcept.strict") + ret <2 x double> %result +} + +define <3 x double> @constrained_vector_fpext_v3f32() { +; PC64LE-LABEL: constrained_vector_fpext_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI101_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI101_1@toc@ha +; PC64LE-NEXT: addis 5, 2, .LCPI101_2@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI101_0@toc@l(3) +; PC64LE-NEXT: lfs 2, .LCPI101_1@toc@l(4) +; PC64LE-NEXT: lfs 3, .LCPI101_2@toc@l(5) +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fpext_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI101_0@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI101_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI101_1@toc@ha +; PC64LE9-NEXT: lfs 2, .LCPI101_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI101_2@toc@ha +; PC64LE9-NEXT: lfs 3, .LCPI101_2@toc@l(3) +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fpext_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI101_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI101_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI101_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI101_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI101_2@toc@ha +; PC64-NEXT: lfs 3, .LCPI101_2@toc@l(3) +; PC64-NEXT: blr +entry: + %result = call <3 x double> @llvm.experimental.constrained.fpext.v3f64.v3f32( + <3 x float>, + metadata !"fpexcept.strict") + ret <3 x double> %result +} + +define <4 x double> @constrained_vector_fpext_v4f32() { +; PC64LE-LABEL: constrained_vector_fpext_v4f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI102_0@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI102_1@toc@ha +; PC64LE-NEXT: addis 5, 2, .LCPI102_2@toc@ha +; PC64LE-NEXT: addis 6, 2, .LCPI102_3@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI102_0@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI102_1@toc@l(4) +; PC64LE-NEXT: lfs 2, .LCPI102_2@toc@l(5) +; PC64LE-NEXT: lfs 3, .LCPI102_3@toc@l(6) +; PC64LE-NEXT: xxmrghd 34, 1, 0 +; PC64LE-NEXT: xxmrghd 35, 3, 2 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_fpext_v4f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI102_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI102_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI102_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI102_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI102_2@toc@ha +; PC64LE9-NEXT: xxmrghd 34, 1, 0 +; PC64LE9-NEXT: lfs 0, .LCPI102_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI102_3@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI102_3@toc@l(3) +; PC64LE9-NEXT: xxmrghd 35, 1, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_fpext_v4f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: addis 3, 2, .LCPI102_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI102_0@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI102_1@toc@ha +; PC64-NEXT: lfs 2, .LCPI102_1@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI102_2@toc@ha +; PC64-NEXT: lfs 3, .LCPI102_2@toc@l(3) +; PC64-NEXT: addis 3, 2, .LCPI102_3@toc@ha +; PC64-NEXT: lfs 4, .LCPI102_3@toc@l(3) +; PC64-NEXT: blr +entry: + %result = call <4 x double> @llvm.experimental.constrained.fpext.v4f64.v4f32( + <4 x float>, + metadata !"fpexcept.strict") + ret <4 x double> %result +} + +define <1 x float> @constrained_vector_ceil_v1f32() { +; PC64LE-LABEL: constrained_vector_ceil_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI103_0@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI103_0@toc@l(3) +; PC64LE-NEXT: frip 0, 0 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_ceil_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI103_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI103_0@toc@l(3) +; PC64LE9-NEXT: frip 0, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_ceil_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI103_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI103_0@toc@l(3) +; PC64-NEXT: bl ceilf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %ceil = call <1 x float> @llvm.experimental.constrained.ceil.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %ceil +} + +define <2 x double> @constrained_vector_ceil_v2f64() { +; PC64LE-LABEL: constrained_vector_ceil_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI104_0@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI104_0@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xvrdpip 34, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_ceil_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI104_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI104_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvrdpip 34, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_ceil_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI104_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI104_0@toc@l(3) +; PC64-NEXT: bl ceil +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI104_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI104_1@toc@l(3) +; PC64-NEXT: bl ceil +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %ceil = call <2 x double> @llvm.experimental.constrained.ceil.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %ceil +} + +define <3 x float> @constrained_vector_ceil_v3f32() { +; PC64LE-LABEL: constrained_vector_ceil_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI105_2@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI105_1@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI105_2@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI105_1@toc@l(4) +; PC64LE-NEXT: addis 3, 2, .LCPI105_0@toc@ha +; PC64LE-NEXT: frip 0, 0 +; PC64LE-NEXT: lfs 2, .LCPI105_0@toc@l(3) +; PC64LE-NEXT: addis 3, 2, .LCPI105_3@toc@ha +; PC64LE-NEXT: frip 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI105_3@toc@l +; PC64LE-NEXT: frip 2, 2 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 2 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 3, 2 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_ceil_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI105_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI105_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI105_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI105_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI105_2@toc@ha +; PC64LE9-NEXT: frip 0, 0 +; PC64LE9-NEXT: lfs 2, .LCPI105_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI105_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI105_3@toc@l +; PC64LE9-NEXT: frip 1, 1 +; PC64LE9-NEXT: frip 2, 2 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xscvdpspn 1, 1 +; PC64LE9-NEXT: xscvdpspn 2, 2 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE9-NEXT: xxsldwi 34, 2, 2, 1 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_ceil_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI105_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI105_0@toc@l(3) +; PC64-NEXT: bl ceilf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI105_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI105_1@toc@l(3) +; PC64-NEXT: bl ceilf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI105_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI105_2@toc@l(3) +; PC64-NEXT: bl ceilf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %ceil = call <3 x float> @llvm.experimental.constrained.ceil.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %ceil +} + +define <3 x double> @constrained_vector_ceil_v3f64() { +; PC64LE-LABEL: constrained_vector_ceil_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI106_1@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI106_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: addis 3, 2, .LCPI106_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI106_0@toc@l(3) +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xsrdpip 3, 1 +; PC64LE-NEXT: xvrdpip 2, 0 +; PC64LE-NEXT: xxswapd 1, 2 +; PC64LE-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_ceil_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI106_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI106_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI106_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI106_1@toc@l +; PC64LE9-NEXT: xsrdpip 3, 0 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvrdpip 2, 0 +; PC64LE9-NEXT: xxswapd 1, 2 +; PC64LE9-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE9-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_ceil_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI106_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI106_0@toc@l(3) +; PC64-NEXT: bl ceil +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI106_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI106_1@toc@l(3) +; PC64-NEXT: bl ceil +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI106_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI106_2@toc@l(3) +; PC64-NEXT: bl ceil +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %ceil = call <3 x double> @llvm.experimental.constrained.ceil.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %ceil +} + +define <1 x float> @constrained_vector_floor_v1f32() { +; PC64LE-LABEL: constrained_vector_floor_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI107_0@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI107_0@toc@l(3) +; PC64LE-NEXT: frim 0, 0 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_floor_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI107_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI107_0@toc@l(3) +; PC64LE9-NEXT: frim 0, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_floor_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI107_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI107_0@toc@l(3) +; PC64-NEXT: bl floorf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %floor = call <1 x float> @llvm.experimental.constrained.floor.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %floor +} + + +define <2 x double> @constrained_vector_floor_v2f64() { +; PC64LE-LABEL: constrained_vector_floor_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI108_0@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI108_0@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xvrdpim 34, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_floor_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI108_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI108_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvrdpim 34, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_floor_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI108_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI108_0@toc@l(3) +; PC64-NEXT: bl floor +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI108_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI108_1@toc@l(3) +; PC64-NEXT: bl floor +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %floor = call <2 x double> @llvm.experimental.constrained.floor.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %floor +} + +define <3 x float> @constrained_vector_floor_v3f32() { +; PC64LE-LABEL: constrained_vector_floor_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI109_2@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI109_1@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI109_2@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI109_1@toc@l(4) +; PC64LE-NEXT: addis 3, 2, .LCPI109_0@toc@ha +; PC64LE-NEXT: frim 0, 0 +; PC64LE-NEXT: lfs 2, .LCPI109_0@toc@l(3) +; PC64LE-NEXT: addis 3, 2, .LCPI109_3@toc@ha +; PC64LE-NEXT: frim 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI109_3@toc@l +; PC64LE-NEXT: frim 2, 2 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 2 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 3, 2 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_floor_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI109_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI109_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI109_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI109_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI109_2@toc@ha +; PC64LE9-NEXT: frim 0, 0 +; PC64LE9-NEXT: lfs 2, .LCPI109_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI109_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI109_3@toc@l +; PC64LE9-NEXT: frim 1, 1 +; PC64LE9-NEXT: frim 2, 2 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xscvdpspn 1, 1 +; PC64LE9-NEXT: xscvdpspn 2, 2 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE9-NEXT: xxsldwi 34, 2, 2, 1 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_floor_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI109_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI109_0@toc@l(3) +; PC64-NEXT: bl floorf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI109_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI109_1@toc@l(3) +; PC64-NEXT: bl floorf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI109_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI109_2@toc@l(3) +; PC64-NEXT: bl floorf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %floor = call <3 x float> @llvm.experimental.constrained.floor.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %floor +} + +define <3 x double> @constrained_vector_floor_v3f64() { +; PC64LE-LABEL: constrained_vector_floor_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI110_1@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI110_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: addis 3, 2, .LCPI110_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI110_0@toc@l(3) +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xsrdpim 3, 1 +; PC64LE-NEXT: xvrdpim 2, 0 +; PC64LE-NEXT: xxswapd 1, 2 +; PC64LE-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_floor_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI110_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI110_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI110_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI110_1@toc@l +; PC64LE9-NEXT: xsrdpim 3, 0 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvrdpim 2, 0 +; PC64LE9-NEXT: xxswapd 1, 2 +; PC64LE9-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE9-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_floor_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI110_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI110_0@toc@l(3) +; PC64-NEXT: bl floor +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI110_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI110_1@toc@l(3) +; PC64-NEXT: bl floor +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI110_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI110_2@toc@l(3) +; PC64-NEXT: bl floor +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %floor = call <3 x double> @llvm.experimental.constrained.floor.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %floor +} + +define <1 x float> @constrained_vector_round_v1f32() { +; PC64LE-LABEL: constrained_vector_round_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI111_0@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI111_0@toc@l(3) +; PC64LE-NEXT: frin 0, 0 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_round_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI111_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI111_0@toc@l(3) +; PC64LE9-NEXT: frin 0, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_round_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI111_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI111_0@toc@l(3) +; PC64-NEXT: bl roundf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %round = call <1 x float> @llvm.experimental.constrained.round.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %round +} + +define <2 x double> @constrained_vector_round_v2f64() { +; PC64LE-LABEL: constrained_vector_round_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI112_0@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI112_0@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xvrdpi 34, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_round_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI112_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI112_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvrdpi 34, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_round_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI112_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI112_0@toc@l(3) +; PC64-NEXT: bl round +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI112_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI112_1@toc@l(3) +; PC64-NEXT: bl round +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %round = call <2 x double> @llvm.experimental.constrained.round.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %round +} + +define <3 x float> @constrained_vector_round_v3f32() { +; PC64LE-LABEL: constrained_vector_round_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI113_2@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI113_1@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI113_2@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI113_1@toc@l(4) +; PC64LE-NEXT: addis 3, 2, .LCPI113_0@toc@ha +; PC64LE-NEXT: frin 0, 0 +; PC64LE-NEXT: lfs 2, .LCPI113_0@toc@l(3) +; PC64LE-NEXT: addis 3, 2, .LCPI113_3@toc@ha +; PC64LE-NEXT: frin 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI113_3@toc@l +; PC64LE-NEXT: frin 2, 2 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 2 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 3, 2 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_round_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI113_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI113_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI113_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI113_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI113_2@toc@ha +; PC64LE9-NEXT: frin 0, 0 +; PC64LE9-NEXT: lfs 2, .LCPI113_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI113_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI113_3@toc@l +; PC64LE9-NEXT: frin 1, 1 +; PC64LE9-NEXT: frin 2, 2 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xscvdpspn 1, 1 +; PC64LE9-NEXT: xscvdpspn 2, 2 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE9-NEXT: xxsldwi 34, 2, 2, 1 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_round_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI113_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI113_0@toc@l(3) +; PC64-NEXT: bl roundf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI113_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI113_1@toc@l(3) +; PC64-NEXT: bl roundf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI113_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI113_2@toc@l(3) +; PC64-NEXT: bl roundf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %round = call <3 x float> @llvm.experimental.constrained.round.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %round +} + + +define <3 x double> @constrained_vector_round_v3f64() { +; PC64LE-LABEL: constrained_vector_round_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI114_1@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI114_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: addis 3, 2, .LCPI114_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI114_0@toc@l(3) +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xsrdpi 3, 1 +; PC64LE-NEXT: xvrdpi 2, 0 +; PC64LE-NEXT: xxswapd 1, 2 +; PC64LE-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_round_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI114_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI114_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI114_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI114_1@toc@l +; PC64LE9-NEXT: xsrdpi 3, 0 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvrdpi 2, 0 +; PC64LE9-NEXT: xxswapd 1, 2 +; PC64LE9-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE9-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_round_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI114_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI114_0@toc@l(3) +; PC64-NEXT: bl round +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI114_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI114_1@toc@l(3) +; PC64-NEXT: bl round +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI114_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI114_2@toc@l(3) +; PC64-NEXT: bl round +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %round = call <3 x double> @llvm.experimental.constrained.round.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %round +} + +define <1 x float> @constrained_vector_trunc_v1f32() { +; PC64LE-LABEL: constrained_vector_trunc_v1f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI115_0@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI115_0@toc@l(3) +; PC64LE-NEXT: friz 0, 0 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_trunc_v1f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI115_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI115_0@toc@l(3) +; PC64LE9-NEXT: friz 0, 0 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_trunc_v1f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -112(1) +; PC64-NEXT: .cfi_def_cfa_offset 112 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: addis 3, 2, .LCPI115_0@toc@ha +; PC64-NEXT: lfs 1, .LCPI115_0@toc@l(3) +; PC64-NEXT: bl truncf +; PC64-NEXT: nop +; PC64-NEXT: addi 1, 1, 112 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %trunc = call <1 x float> @llvm.experimental.constrained.trunc.v1f32( + <1 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <1 x float> %trunc +} + +define <2 x double> @constrained_vector_trunc_v2f64() { +; PC64LE-LABEL: constrained_vector_trunc_v2f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI116_0@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI116_0@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xvrdpiz 34, 0 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_trunc_v2f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI116_0@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI116_0@toc@l +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvrdpiz 34, 0 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_trunc_v2f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI116_0@toc@ha +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI116_0@toc@l(3) +; PC64-NEXT: bl trunc +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI116_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI116_1@toc@l(3) +; PC64-NEXT: bl trunc +; PC64-NEXT: nop +; PC64-NEXT: fmr 2, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %trunc = call <2 x double> @llvm.experimental.constrained.trunc.v2f64( + <2 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <2 x double> %trunc +} + +define <3 x float> @constrained_vector_trunc_v3f32() { +; PC64LE-LABEL: constrained_vector_trunc_v3f32: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI117_2@toc@ha +; PC64LE-NEXT: addis 4, 2, .LCPI117_1@toc@ha +; PC64LE-NEXT: lfs 0, .LCPI117_2@toc@l(3) +; PC64LE-NEXT: lfs 1, .LCPI117_1@toc@l(4) +; PC64LE-NEXT: addis 3, 2, .LCPI117_0@toc@ha +; PC64LE-NEXT: friz 0, 0 +; PC64LE-NEXT: lfs 2, .LCPI117_0@toc@l(3) +; PC64LE-NEXT: addis 3, 2, .LCPI117_3@toc@ha +; PC64LE-NEXT: friz 1, 1 +; PC64LE-NEXT: addi 3, 3, .LCPI117_3@toc@l +; PC64LE-NEXT: friz 2, 2 +; PC64LE-NEXT: xscvdpspn 0, 0 +; PC64LE-NEXT: xscvdpspn 1, 1 +; PC64LE-NEXT: xxsldwi 34, 0, 0, 1 +; PC64LE-NEXT: xscvdpspn 0, 2 +; PC64LE-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE-NEXT: vmrglw 2, 3, 2 +; PC64LE-NEXT: lvx 3, 0, 3 +; PC64LE-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE-NEXT: vperm 2, 4, 2, 3 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_trunc_v3f32: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI117_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI117_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI117_1@toc@ha +; PC64LE9-NEXT: lfs 1, .LCPI117_1@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI117_2@toc@ha +; PC64LE9-NEXT: friz 0, 0 +; PC64LE9-NEXT: lfs 2, .LCPI117_2@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI117_3@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI117_3@toc@l +; PC64LE9-NEXT: friz 1, 1 +; PC64LE9-NEXT: friz 2, 2 +; PC64LE9-NEXT: xscvdpspn 0, 0 +; PC64LE9-NEXT: xscvdpspn 1, 1 +; PC64LE9-NEXT: xscvdpspn 2, 2 +; PC64LE9-NEXT: xxsldwi 36, 0, 0, 1 +; PC64LE9-NEXT: xxsldwi 35, 1, 1, 1 +; PC64LE9-NEXT: xxsldwi 34, 2, 2, 1 +; PC64LE9-NEXT: vmrglw 2, 3, 2 +; PC64LE9-NEXT: lxvx 35, 0, 3 +; PC64LE9-NEXT: vperm 2, 4, 2, 3 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_trunc_v3f32: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI117_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfs 1, .LCPI117_0@toc@l(3) +; PC64-NEXT: bl truncf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI117_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfs 1, .LCPI117_1@toc@l(3) +; PC64-NEXT: bl truncf +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI117_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI117_2@toc@l(3) +; PC64-NEXT: bl truncf +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %trunc = call <3 x float> @llvm.experimental.constrained.trunc.v3f32( + <3 x float> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x float> %trunc +} + +define <3 x double> @constrained_vector_trunc_v3f64() { +; PC64LE-LABEL: constrained_vector_trunc_v3f64: +; PC64LE: # %bb.0: # %entry +; PC64LE-NEXT: addis 3, 2, .LCPI118_1@toc@ha +; PC64LE-NEXT: addi 3, 3, .LCPI118_1@toc@l +; PC64LE-NEXT: lxvd2x 0, 0, 3 +; PC64LE-NEXT: addis 3, 2, .LCPI118_0@toc@ha +; PC64LE-NEXT: lfs 1, .LCPI118_0@toc@l(3) +; PC64LE-NEXT: xxswapd 0, 0 +; PC64LE-NEXT: xsrdpiz 3, 1 +; PC64LE-NEXT: xvrdpiz 2, 0 +; PC64LE-NEXT: xxswapd 1, 2 +; PC64LE-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE-NEXT: blr +; +; PC64LE9-LABEL: constrained_vector_trunc_v3f64: +; PC64LE9: # %bb.0: # %entry +; PC64LE9-NEXT: addis 3, 2, .LCPI118_0@toc@ha +; PC64LE9-NEXT: lfs 0, .LCPI118_0@toc@l(3) +; PC64LE9-NEXT: addis 3, 2, .LCPI118_1@toc@ha +; PC64LE9-NEXT: addi 3, 3, .LCPI118_1@toc@l +; PC64LE9-NEXT: xsrdpiz 3, 0 +; PC64LE9-NEXT: lxvx 0, 0, 3 +; PC64LE9-NEXT: xvrdpiz 2, 0 +; PC64LE9-NEXT: xxswapd 1, 2 +; PC64LE9-NEXT: # kill: def $f1 killed $f1 killed $vsl1 +; PC64LE9-NEXT: # kill: def $f2 killed $f2 killed $vsl2 +; PC64LE9-NEXT: blr +; +; PC64-LABEL: constrained_vector_trunc_v3f64: +; PC64: # %bb.0: # %entry +; PC64-NEXT: mflr 0 +; PC64-NEXT: std 0, 16(1) +; PC64-NEXT: stdu 1, -128(1) +; PC64-NEXT: .cfi_def_cfa_offset 128 +; PC64-NEXT: .cfi_offset lr, 16 +; PC64-NEXT: .cfi_offset f30, -16 +; PC64-NEXT: .cfi_offset f31, -8 +; PC64-NEXT: addis 3, 2, .LCPI118_0@toc@ha +; PC64-NEXT: stfd 30, 112(1) # 8-byte Folded Spill +; PC64-NEXT: stfd 31, 120(1) # 8-byte Folded Spill +; PC64-NEXT: lfd 1, .LCPI118_0@toc@l(3) +; PC64-NEXT: bl trunc +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI118_1@toc@ha +; PC64-NEXT: fmr 31, 1 +; PC64-NEXT: lfd 1, .LCPI118_1@toc@l(3) +; PC64-NEXT: bl trunc +; PC64-NEXT: nop +; PC64-NEXT: addis 3, 2, .LCPI118_2@toc@ha +; PC64-NEXT: fmr 30, 1 +; PC64-NEXT: lfs 1, .LCPI118_2@toc@l(3) +; PC64-NEXT: bl trunc +; PC64-NEXT: nop +; PC64-NEXT: fmr 3, 1 +; PC64-NEXT: fmr 1, 31 +; PC64-NEXT: fmr 2, 30 +; PC64-NEXT: lfd 31, 120(1) # 8-byte Folded Reload +; PC64-NEXT: lfd 30, 112(1) # 8-byte Folded Reload +; PC64-NEXT: addi 1, 1, 128 +; PC64-NEXT: ld 0, 16(1) +; PC64-NEXT: mtlr 0 +; PC64-NEXT: blr +entry: + %trunc = call <3 x double> @llvm.experimental.constrained.trunc.v3f64( + <3 x double> , + metadata !"round.dynamic", + metadata !"fpexcept.strict") + ret <3 x double> %trunc +} + + +; Single width declarations +declare <2 x double> @llvm.experimental.constrained.fadd.v2f64(<2 x double>, <2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.fsub.v2f64(<2 x double>, <2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.fmul.v2f64(<2 x double>, <2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.fdiv.v2f64(<2 x double>, <2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.frem.v2f64(<2 x double>, <2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.sqrt.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.pow.v2f64(<2 x double>, <2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.powi.v2f64(<2 x double>, i32, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.sin.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.cos.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.exp.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.exp2.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.log.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.log10.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.log2.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.rint.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.nearbyint.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.maxnum.v2f64(<2 x double>, <2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.minnum.v2f64(<2 x double>, <2 x double>, metadata, metadata) +declare <2 x float> @llvm.experimental.constrained.fptrunc.v2f32.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.fpext.v2f64.v2f32(<2 x float>, metadata) +declare <2 x double> @llvm.experimental.constrained.ceil.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.floor.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.round.v2f64(<2 x double>, metadata, metadata) +declare <2 x double> @llvm.experimental.constrained.trunc.v2f64(<2 x double>, metadata, metadata) + +; Scalar width declarations +declare <1 x float> @llvm.experimental.constrained.fadd.v1f32(<1 x float>, <1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.fsub.v1f32(<1 x float>, <1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.fmul.v1f32(<1 x float>, <1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.fdiv.v1f32(<1 x float>, <1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.frem.v1f32(<1 x float>, <1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.sqrt.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.pow.v1f32(<1 x float>, <1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.powi.v1f32(<1 x float>, i32, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.sin.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.cos.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.exp.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.exp2.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.log.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.log10.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.log2.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.rint.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.nearbyint.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.maxnum.v1f32(<1 x float>, <1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.minnum.v1f32(<1 x float>, <1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.fptrunc.v1f32.v1f64(<1 x double>, metadata, metadata) +declare <1 x double> @llvm.experimental.constrained.fpext.v1f64.v1f32(<1 x float>, metadata) +declare <1 x float> @llvm.experimental.constrained.ceil.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.floor.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.round.v1f32(<1 x float>, metadata, metadata) +declare <1 x float> @llvm.experimental.constrained.trunc.v1f32(<1 x float>, metadata, metadata) + +; Illegal width declarations +declare <3 x float> @llvm.experimental.constrained.fadd.v3f32(<3 x float>, <3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.fadd.v3f64(<3 x double>, <3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.fsub.v3f32(<3 x float>, <3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.fsub.v3f64(<3 x double>, <3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.fmul.v3f32(<3 x float>, <3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.fmul.v3f64(<3 x double>, <3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.fdiv.v3f32(<3 x float>, <3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.fdiv.v3f64(<3 x double>, <3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.frem.v3f32(<3 x float>, <3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.frem.v3f64(<3 x double>, <3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.sqrt.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.sqrt.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.pow.v3f32(<3 x float>, <3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.pow.v3f64(<3 x double>, <3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.powi.v3f32(<3 x float>, i32, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.powi.v3f64(<3 x double>, i32, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.sin.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.sin.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.cos.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.cos.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.exp.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.exp.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.exp2.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.exp2.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.log.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.log.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.log10.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.log10.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.log2.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.log2.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.rint.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.rint.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.nearbyint.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.nearbyint.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.maxnum.v3f32(<3 x float>, <3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.maxnum.v3f64(<3 x double>, <3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.minnum.v3f32(<3 x float>, <3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.minnum.v3f64(<3 x double>, <3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.fptrunc.v3f32.v3f64(<3 x double>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.fpext.v3f64.v3f32(<3 x float>, metadata) +declare <3 x float> @llvm.experimental.constrained.ceil.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.ceil.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.floor.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.floor.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.round.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.round.v3f64(<3 x double>, metadata, metadata) +declare <3 x float> @llvm.experimental.constrained.trunc.v3f32(<3 x float>, metadata, metadata) +declare <3 x double> @llvm.experimental.constrained.trunc.v3f64(<3 x double>, metadata, metadata) + +; Double width declarations +declare <4 x double> @llvm.experimental.constrained.fadd.v4f64(<4 x double>, <4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.fsub.v4f64(<4 x double>, <4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.fmul.v4f64(<4 x double>, <4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.fdiv.v4f64(<4 x double>, <4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.frem.v4f64(<4 x double>, <4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.sqrt.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.pow.v4f64(<4 x double>, <4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.powi.v4f64(<4 x double>, i32, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.sin.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.cos.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.exp.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.exp2.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.log.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.log10.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.log2.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.rint.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.nearbyint.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.maxnum.v4f64(<4 x double>, <4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.minnum.v4f64(<4 x double>, <4 x double>, metadata, metadata) +declare <4 x float> @llvm.experimental.constrained.fptrunc.v4f32.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.fpext.v4f64.v4f32(<4 x float>, metadata) +declare <4 x double> @llvm.experimental.constrained.ceil.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.floor.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.round.v4f64(<4 x double>, metadata, metadata) +declare <4 x double> @llvm.experimental.constrained.trunc.v4f64(<4 x double>, metadata, metadata)