1; RUN: llc < %s -march=x86 -mcpu=core2 | FileCheck %s 2; RUN: llc < %s -march=x86-64 -mcpu=corei7 | FileCheck %s 3 4; Verify that the backend correctly combines SSE2 builtin intrinsics. 5 6 7define <4 x i32> @test_psra_1(<4 x i32> %A) { 8 %1 = tail call <4 x i32> @llvm.x86.sse2.psrai.d(<4 x i32> %A, i32 3) 9 %2 = tail call <4 x i32> @llvm.x86.sse2.psra.d(<4 x i32> %1, <4 x i32> <i32 3, i32 0, i32 7, i32 0>) 10 %3 = tail call <4 x i32> @llvm.x86.sse2.psrai.d(<4 x i32> %2, i32 2) 11 ret <4 x i32> %3 12} 13; CHECK-LABEL: test_psra_1 14; CHECK: psrad $8, %xmm0 15; CHECK-NEXT: ret 16 17define <8 x i16> @test_psra_2(<8 x i16> %A) { 18 %1 = tail call <8 x i16> @llvm.x86.sse2.psrai.w(<8 x i16> %A, i32 3) 19 %2 = tail call <8 x i16> @llvm.x86.sse2.psra.w(<8 x i16> %1, <8 x i16> <i16 3, i16 0, i16 0, i16 0, i16 7, i16 0, i16 0, i16 0>) 20 %3 = tail call <8 x i16> @llvm.x86.sse2.psrai.w(<8 x i16> %2, i32 2) 21 ret <8 x i16> %3 22} 23; CHECK-LABEL: test_psra_2 24; CHECK: psraw $8, %xmm0 25; CHECK-NEXT: ret 26 27define <4 x i32> @test_psra_3(<4 x i32> %A) { 28 %1 = tail call <4 x i32> @llvm.x86.sse2.psrai.d(<4 x i32> %A, i32 0) 29 %2 = tail call <4 x i32> @llvm.x86.sse2.psra.d(<4 x i32> %1, <4 x i32> <i32 0, i32 0, i32 7, i32 0>) 30 %3 = tail call <4 x i32> @llvm.x86.sse2.psrai.d(<4 x i32> %2, i32 0) 31 ret <4 x i32> %3 32} 33; CHECK-LABEL: test_psra_3 34; CHECK-NOT: psrad 35; CHECK: ret 36 37 38define <8 x i16> @test_psra_4(<8 x i16> %A) { 39 %1 = tail call <8 x i16> @llvm.x86.sse2.psrai.w(<8 x i16> %A, i32 0) 40 %2 = tail call <8 x i16> @llvm.x86.sse2.psra.w(<8 x i16> %1, <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 7, i16 0, i16 0, i16 0>) 41 %3 = tail call <8 x i16> @llvm.x86.sse2.psrai.w(<8 x i16> %2, i32 0) 42 ret <8 x i16> %3 43} 44; CHECK-LABEL: test_psra_4 45; CHECK-NOT: psraw 46; CHECK: ret 47 48 49declare <8 x i16> @llvm.x86.sse2.psra.w(<8 x i16>, <8 x i16>) 50declare <8 x i16> @llvm.x86.sse2.psrai.w(<8 x i16>, i32) 51declare <4 x i32> @llvm.x86.sse2.psra.d(<4 x i32>, <4 x i32>) 52declare <4 x i32> @llvm.x86.sse2.psrai.d(<4 x i32>, i32) 53 54