[RISCV][CodeGen][NFC] Add fixed vector type test cases for llvm.round.*

[frasercrmck]

llvm/test/CodeGen/RISCV/rvv/fround-sdnode.ll

297

These should probably be in fixed-vectors-round.ll to fit in with our other tests.

[eopXD]

Follow current convention and sort test cases correctly.

[eopXD]

llvm/test/CodeGen/RISCV/rvv/fround-sdnode.ll

297

Thank you for pointing this out.

[eopXD]

Probably should rename current test cases of xxx-sdnode.ll to xxx.ll too to make file naming more consistent.

[reames]

LGTM

[craig.topper]

What was wrong with the tests in fixed-vectors-fp.ll other than it not covering as many types?

[eopXD]

In D134799#3821276, @craig.topper wrote:

What was wrong with the tests in fixed-vectors-fp.ll other than it not covering as many types?

Sorry that I didn't notice test cases where collected under fixed-vectors-fp.ll. Do you recommend to revert this commit?

[craig.topper]

In D134799#3821287, @eopXD wrote:

In D134799#3821276, @craig.topper wrote:

What was wrong with the tests in fixed-vectors-fp.ll other than it not covering as many types?

Sorry that I didn't notice test cases where collected under fixed-vectors-fp.ll. Do you recommend to revert this commit?

Don't revert it. We should probably remove the redundant round tests from fixed-vectors-fp.ll

llvm/test/CodeGen/RISCV/rvv/fround-sdnode.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -mtriple=riscv32 -mattr=+d,+zfh,+experimental-zvfh,+v -target-abi=ilp32d \
 ; RUN:     -verify-machineinstrs < %s | FileCheck %s
 ; RUN: llc -mtriple=riscv64 -mattr=+d,+zfh,+experimental-zvfh,+v -target-abi=lp64d \
 ; RUN:     -verify-machineinstrs < %s | FileCheck %s
 
+; This file tests the code generation for `llvm.round.*`.
+; Scalable vector type test cases comes first, then fixed vector types.
+
+; Scalable vector type
 define <vscale x 1 x half> @round_nxv1f16(<vscale x 1 x half> %x) {
 ; CHECK-LABEL: round_nxv1f16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    lui a0, %hi(.LCPI0_0)
 ; CHECK-NEXT:    flh ft0, %lo(.LCPI0_0)(a0)
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, mu
 ; CHECK-NEXT:    vfabs.v v9, v8
 ; CHECK-NEXT:    vmflt.vf v0, v9, ft0

 ; CHECK-NEXT:    fsrm a0
 ; CHECK-NEXT:    vfcvt.f.x.v v16, v16, v0.t
 ; CHECK-NEXT:    vfsgnj.vv v8, v16, v8, v0.t
 ; CHECK-NEXT:    ret
   %a = call <vscale x 8 x double> @llvm.round.nxv8f64(<vscale x 8 x double> %x)
   ret <vscale x 8 x double> %a
 }
 declare <vscale x 8 x double> @llvm.round.nxv8f64(<vscale x 8 x double>)
+
+; Fixed vector type
+define <1 x half> @round_v1f16(<1 x half> %x) {
+; CHECK-LABEL: round_v1f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI15_0)
+; CHECK-NEXT:    flh ft0, %lo(.LCPI15_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 1, e16, mf4, ta, mu
+; CHECK-NEXT:    vfabs.v v9, v8
+; CHECK-NEXT:    vmflt.vf v0, v9, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v9, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v9, v9, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v9, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <1 x half> @llvm.round.v1f16(<1 x half> %x)
+  ret <1 x half> %a
+}
+declare <1 x half> @llvm.round.v1f16(<1 x half>)
+
+define <2 x half> @round_v2f16(<2 x half> %x) {
+; CHECK-LABEL: round_v2f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI16_0)
+; CHECK-NEXT:    flh ft0, %lo(.LCPI16_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, mu
+; CHECK-NEXT:    vfabs.v v9, v8
+; CHECK-NEXT:    vmflt.vf v0, v9, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v9, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v9, v9, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v9, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <2 x half> @llvm.round.v2f16(<2 x half> %x)
+  ret <2 x half> %a
+}
+declare <2 x half> @llvm.round.v2f16(<2 x half>)
+
+define <4 x half> @round_v4f16(<4 x half> %x) {
+; CHECK-LABEL: round_v4f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI17_0)
+; CHECK-NEXT:    flh ft0, %lo(.LCPI17_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, mu
+; CHECK-NEXT:    vfabs.v v9, v8
+; CHECK-NEXT:    vmflt.vf v0, v9, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v9, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v9, v9, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v9, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <4 x half> @llvm.round.v4f16(<4 x half> %x)
+  ret <4 x half> %a
+}
+declare <4 x half> @llvm.round.v4f16(<4 x half>)
+
+define <8 x half> @round_v8f16(<8 x half> %x) {
+; CHECK-LABEL: round_v8f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI18_0)
+; CHECK-NEXT:    flh ft0, %lo(.LCPI18_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, mu
+; CHECK-NEXT:    vfabs.v v9, v8
+; CHECK-NEXT:    vmflt.vf v0, v9, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v9, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v9, v9, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v9, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <8 x half> @llvm.round.v8f16(<8 x half> %x)
+  ret <8 x half> %a
+}
+declare <8 x half> @llvm.round.v8f16(<8 x half>)
+
+define <16 x half> @round_v16f16(<16 x half> %x) {
+; CHECK-LABEL: round_v16f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI19_0)
+; CHECK-NEXT:    flh ft0, %lo(.LCPI19_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, mu
+; CHECK-NEXT:    vfabs.v v10, v8
+; CHECK-NEXT:    vmflt.vf v0, v10, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v10, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v10, v10, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v10, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <16 x half> @llvm.round.v16f16(<16 x half> %x)
+  ret <16 x half> %a
+}
+declare <16 x half> @llvm.round.v16f16(<16 x half>)
+
+define <32 x half> @round_v32f16(<32 x half> %x) {
+; CHECK-LABEL: round_v32f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI20_0)
+; CHECK-NEXT:    flh ft0, %lo(.LCPI20_0)(a0)
+; CHECK-NEXT:    li a0, 32
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, mu
+; CHECK-NEXT:    vfabs.v v12, v8
+; CHECK-NEXT:    vmflt.vf v0, v12, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v12, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v12, v12, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v12, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <32 x half> @llvm.round.v32f16(<32 x half> %x)
+  ret <32 x half> %a
+}
+declare <32 x half> @llvm.round.v32f16(<32 x half>)
+
+define <1 x float> @round_v1f32(<1 x float> %x) {
+; CHECK-LABEL: round_v1f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI21_0)
+; CHECK-NEXT:    flw ft0, %lo(.LCPI21_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 1, e32, mf2, ta, mu
+; CHECK-NEXT:    vfabs.v v9, v8
+; CHECK-NEXT:    vmflt.vf v0, v9, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v9, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v9, v9, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v9, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <1 x float> @llvm.round.v1f32(<1 x float> %x)
+  ret <1 x float> %a
+}
+declare <1 x float> @llvm.round.v1f32(<1 x float>)
+
+define <2 x float> @round_v2f32(<2 x float> %x) {
+; CHECK-LABEL: round_v2f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI22_0)
+; CHECK-NEXT:    flw ft0, %lo(.LCPI22_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, mu
+; CHECK-NEXT:    vfabs.v v9, v8
+; CHECK-NEXT:    vmflt.vf v0, v9, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v9, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v9, v9, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v9, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <2 x float> @llvm.round.v2f32(<2 x float> %x)
+  ret <2 x float> %a
+}
+declare <2 x float> @llvm.round.v2f32(<2 x float>)
+
+define <4 x float> @round_v4f32(<4 x float> %x) {
+; CHECK-LABEL: round_v4f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI23_0)
+; CHECK-NEXT:    flw ft0, %lo(.LCPI23_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, mu
+; CHECK-NEXT:    vfabs.v v9, v8
+; CHECK-NEXT:    vmflt.vf v0, v9, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v9, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v9, v9, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v9, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <4 x float> @llvm.round.v4f32(<4 x float> %x)
+  ret <4 x float> %a
+}
+declare <4 x float> @llvm.round.v4f32(<4 x float>)
+
+define <8 x float> @round_v8f32(<8 x float> %x) {
+; CHECK-LABEL: round_v8f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI24_0)
+; CHECK-NEXT:    flw ft0, %lo(.LCPI24_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 8, e32, m2, ta, mu
+; CHECK-NEXT:    vfabs.v v10, v8
+; CHECK-NEXT:    vmflt.vf v0, v10, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v10, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v10, v10, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v10, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <8 x float> @llvm.round.v8f32(<8 x float> %x)
+  ret <8 x float> %a
+}
+declare <8 x float> @llvm.round.v8f32(<8 x float>)
+
+define <16 x float> @round_v16f32(<16 x float> %x) {
+; CHECK-LABEL: round_v16f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI25_0)
+; CHECK-NEXT:    flw ft0, %lo(.LCPI25_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 16, e32, m4, ta, mu
+; CHECK-NEXT:    vfabs.v v12, v8
+; CHECK-NEXT:    vmflt.vf v0, v12, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v12, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v12, v12, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v12, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <16 x float> @llvm.round.v16f32(<16 x float> %x)
+  ret <16 x float> %a
+}
+declare <16 x float> @llvm.round.v16f32(<16 x float>)
+
+define <1 x double> @round_v1f64(<1 x double> %x) {
+; CHECK-LABEL: round_v1f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI26_0)
+; CHECK-NEXT:    fld ft0, %lo(.LCPI26_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 1, e64, m1, ta, mu
+; CHECK-NEXT:    vfabs.v v9, v8
+; CHECK-NEXT:    vmflt.vf v0, v9, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v9, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v9, v9, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v9, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <1 x double> @llvm.round.v1f64(<1 x double> %x)
+  ret <1 x double> %a
+}
+declare <1 x double> @llvm.round.v1f64(<1 x double>)
+
+define <2 x double> @round_v2f64(<2 x double> %x) {
+; CHECK-LABEL: round_v2f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI27_0)
+; CHECK-NEXT:    fld ft0, %lo(.LCPI27_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 2, e64, m1, ta, mu
+; CHECK-NEXT:    vfabs.v v9, v8
+; CHECK-NEXT:    vmflt.vf v0, v9, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v9, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v9, v9, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v9, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <2 x double> @llvm.round.v2f64(<2 x double> %x)
+  ret <2 x double> %a
+}
+declare <2 x double> @llvm.round.v2f64(<2 x double>)
+
+define <4 x double> @round_v4f64(<4 x double> %x) {
+; CHECK-LABEL: round_v4f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI28_0)
+; CHECK-NEXT:    fld ft0, %lo(.LCPI28_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 4, e64, m2, ta, mu
+; CHECK-NEXT:    vfabs.v v10, v8
+; CHECK-NEXT:    vmflt.vf v0, v10, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v10, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v10, v10, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v10, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <4 x double> @llvm.round.v4f64(<4 x double> %x)
+  ret <4 x double> %a
+}
+declare <4 x double> @llvm.round.v4f64(<4 x double>)
+
+define <8 x double> @round_v8f64(<8 x double> %x) {
+; CHECK-LABEL: round_v8f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lui a0, %hi(.LCPI29_0)
+; CHECK-NEXT:    fld ft0, %lo(.LCPI29_0)(a0)
+; CHECK-NEXT:    vsetivli zero, 8, e64, m4, ta, mu
+; CHECK-NEXT:    vfabs.v v12, v8
+; CHECK-NEXT:    vmflt.vf v0, v12, ft0
+; CHECK-NEXT:    fsrmi a0, 4
+; CHECK-NEXT:    vfcvt.x.f.v v12, v8, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vfcvt.f.x.v v12, v12, v0.t
+; CHECK-NEXT:    vfsgnj.vv v8, v12, v8, v0.t
+; CHECK-NEXT:    ret
+  %a = call <8 x double> @llvm.round.v8f64(<8 x double> %x)
+  ret <8 x double> %a
+}
+declare <8 x double> @llvm.round.v8f64(<8 x double>)