Index: include/llvm/IR/IntrinsicsPowerPC.td =================================================================== --- include/llvm/IR/IntrinsicsPowerPC.td +++ include/llvm/IR/IntrinsicsPowerPC.td @@ -694,6 +694,12 @@ def int_ppc_vsx_xvrdpip : Intrinsic<[llvm_v2f64_ty], [llvm_v2f64_ty], [IntrNoMem]>; +// Vector rsqrte +def int_ppc_vsx_xvrsqrtesp : GCCBuiltin<"__builtin_vsx_xvrsqrtesp">, + Intrinsic<[llvm_v4f32_ty], [llvm_v4f32_ty], [IntrNoMem]>; +def int_ppc_vsx_xvrsqrtedp : GCCBuiltin<"__builtin_vsx_xvrsqrtedp">, + Intrinsic<[llvm_v2f64_ty], [llvm_v2f64_ty], [IntrNoMem]>; + // Vector compare def int_ppc_vsx_xvcmpeqdp : PowerPC_VSX_Intrinsic<"xvcmpeqdp", [llvm_v2i64_ty], Index: lib/Target/PowerPC/PPCInstrVSX.td =================================================================== --- lib/Target/PowerPC/PPCInstrVSX.td +++ lib/Target/PowerPC/PPCInstrVSX.td @@ -989,6 +989,12 @@ def : Pat<(int_ppc_vsx_xvdivdp v2f64:$A, v2f64:$B), (XVDIVDP $A, $B)>; +// Recip. square root estimate +def : Pat<(int_ppc_vsx_xvrsqrtesp v4f32:$A), + (XVRSQRTESP $A)>; +def : Pat<(int_ppc_vsx_xvrsqrtedp v2f64:$A), + (XVRSQRTEDP $A)>; + } // AddedComplexity } // HasVSX Index: test/CodeGen/PowerPC/vsx-elementary-arith.ll =================================================================== --- test/CodeGen/PowerPC/vsx-elementary-arith.ll +++ test/CodeGen/PowerPC/vsx-elementary-arith.ll @@ -116,5 +116,31 @@ ; CHECK: xssqrtdp {{[0-9]+}} } +; Vector forms +; Function Attrs: nounwind +define void @emit_xvrsqrte() #0 { +entry: +; CHECK-LABEL: @emit_xvrsqrte + %vf = alloca <4 x float>, align 16 + %vfr = alloca <4 x float>, align 16 + %vd = alloca <2 x double>, align 16 + %vdr = alloca <2 x double>, align 16 + %0 = load <4 x float>, <4 x float>* %vf, align 16 + %1 = call <4 x float> @llvm.ppc.vsx.xvrsqrtesp(<4 x float> %0) +; CHECK: xvrsqrtesp {{[0-9]+}}, {{[0-9]+}} + store <4 x float> %1, <4 x float>* %vfr, align 16 + %2 = load <2 x double>, <2 x double>* %vd, align 16 + %3 = call <2 x double> @llvm.ppc.vsx.xvrsqrtedp(<2 x double> %2) +; CHECK: xvrsqrtedp {{[0-9]+}}, {{[0-9]+}} + store <2 x double> %3, <2 x double>* %vdr, align 16 + ret void +} + ; Function Attrs: nounwind declare double @sqrt(double) + +; Function Attrs: nounwind readnone +declare <4 x float> @llvm.ppc.vsx.xvrsqrtesp(<4 x float>) + +; Function Attrs: nounwind readnone +declare <2 x double> @llvm.ppc.vsx.xvrsqrtedp(<2 x double>)