diff options
| -rw-r--r-- | llvm/test/CodeGen/X86/avx2-intrinsics-fast-isel.ll | 697 |
1 files changed, 665 insertions, 32 deletions
diff --git a/llvm/test/CodeGen/X86/avx2-intrinsics-fast-isel.ll b/llvm/test/CodeGen/X86/avx2-intrinsics-fast-isel.ll index 9488889fb40..684412e38e8 100644 --- a/llvm/test/CodeGen/X86/avx2-intrinsics-fast-isel.ll +++ b/llvm/test/CodeGen/X86/avx2-intrinsics-fast-isel.ll @@ -1051,38 +1051,665 @@ define <4 x i64> @test_mm256_hsubs_epi16(<4 x i64> %a0, <4 x i64> %a1) { } declare <16 x i16> @llvm.x86.avx2.phsub.sw(<16 x i16>, <16 x i16>) nounwind readnone -; TODO test_mm_i32gather_epi32 (int const* base_addr, __m128i vindex, const int scale) -; TODO test_mm_mask_i32gather_epi32 (__m128i src, int const* base_addr, __m128i vindex, __m128i mask, const int scale) -; TODO test_mm256_i32gather_epi32 (int const* base_addr, __m256i vindex, const int scale) -; TODO test_mm256_mask_i32gather_epi32 (__m256i src, int const* base_addr, __m256i vindex, __m256i mask, const int scale) -; TODO test_mm_i32gather_epi64 (__int64 const* base_addr, __m128i vindex, const int scale) -; TODO test_mm_mask_i32gather_epi64 (__m128i src, __int64 const* base_addr, __m128i vindex, __m128i mask, const int scale) -; TODO test_mm256_i32gather_epi64 (__int64 const* base_addr, __m128i vindex, const int scale) -; TODO test_mm256_mask_i32gather_epi64 (__m256i src, __int64 const* base_addr, __m128i vindex, __m256i mask, const int scale) -; TODO test_mm_i32gather_pd (double const* base_addr, __m128i vindex, const int scale) -; TODO test_mm_mask_i32gather_pd (__m128d src, double const* base_addr, __m128i vindex, __m128d mask, const int scale) -; TODO test_mm256_i32gather_pd (double const* base_addr, __m128i vindex, const int scale) -; TODO test_mm256_mask_i32gather_pd (__m256d src, double const* base_addr, __m128i vindex, __m256d mask, const int scale) -; TODO test_mm_i32gather_ps (float const* base_addr, __m128i vindex, const int scale) -; TODO test_mm_mask_i32gather_ps (__m128 src, float const* base_addr, __m128i vindex, __m128 mask, const int scale) -; TODO test_mm256_i32gather_ps (float const* base_addr, __m256i vindex, const int scale) -; TODO test_mm256_mask_i32gather_ps (__m256 src, float const* base_addr, __m256i vindex, __m256 mask, const int scale) -; TODO test_mm_i64gather_epi32 (int const* base_addr, __m128i vindex, const int scale) -; TODO test_mm_mask_i64gather_epi32 (__m128i src, int const* base_addr, __m128i vindex, __m128i mask, const int scale) -; TODO test_mm256_i64gather_epi32 (int const* base_addr, __m256i vindex, const int scale) -; TODO test_mm256_mask_i64gather_epi32 (__m128i src, int const* base_addr, __m256i vindex, __m128i mask, const int scale) -; TODO test_mm_i64gather_epi64 (__int64 const* base_addr, __m128i vindex, const int scale) -; TODO test_mm_mask_i64gather_epi64 (__m128i src, __int64 const* base_addr, __m128i vindex, __m128i mask, const int scale) -; TODO test_mm256_i64gather_epi64 (__int64 const* base_addr, __m256i vindex, const int scale) -; TODO test_mm256_mask_i64gather_epi64 (__m256i src, __int64 const* base_addr, __m256i vindex, __m256i mask, const int scale) -; TODO test_mm_i64gather_pd (double const* base_addr, __m128i vindex, const int scale) -; TODO test_mm_mask_i64gather_pd (__m128d src, double const* base_addr, __m128i vindex, __m128d mask, const int scale) -; TODO test_mm256_i64gather_pd (double const* base_addr, __m256i vindex, const int scale) -; TODO test_mm256_mask_i64gather_pd (__m256d src, double const* base_addr, __m256i vindex, __m256d mask, const int scale) -; TODO test_mm_i64gather_ps (float const* base_addr, __m128i vindex, const int scale) -; TODO test_mm_mask_i64gather_ps (__m128 src, float const* base_addr, __m128i vindex, __m128 mask, const int scale) -; TODO test_mm256_i64gather_ps (float const* base_addr, __m256i vindex, const int scale) -; TODO test_mm256_mask_i64gather_ps (__m128 src, float const* base_addr, __m256i vindex, __m128 mask, const int scale) +define <2 x i64> @test_mm_i32gather_epi32(i32 *%a0, <2 x i64> %a1) { +; X32-LABEL: test_mm_i32gather_epi32: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 +; X32-NEXT: vpgatherdd %xmm2, (%eax,%xmm0,2), %xmm1 +; X32-NEXT: vmovdqa %xmm1, %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_i32gather_epi32: +; X64: # BB#0: +; X64-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 +; X64-NEXT: vpgatherdd %xmm2, (%rdi,%xmm0,2), %xmm1 +; X64-NEXT: vmovdqa %xmm1, %xmm0 +; X64-NEXT: retq + %arg0 = bitcast i32 *%a0 to i8* + %arg1 = bitcast <2 x i64> %a1 to <4 x i32> + %mask = bitcast <2 x i64> <i64 -1, i64 -1> to <4 x i32> + %call = call <4 x i32> @llvm.x86.avx2.gather.d.d(<4 x i32> undef, i8* %arg0, <4 x i32> %arg1, <4 x i32> %mask, i8 2) + %bc = bitcast <4 x i32> %call to <2 x i64> + ret <2 x i64> %bc +} +declare <4 x i32> @llvm.x86.avx2.gather.d.d(<4 x i32>, i8*, <4 x i32>, <4 x i32>, i8) nounwind readonly + +define <2 x i64> @test_mm_mask_i32gather_epi32(<2 x i64> %a0, i32 *%a1, <2 x i64> %a2, <2 x i64> %a3) { +; X32-LABEL: test_mm_mask_i32gather_epi32: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpgatherdd %xmm2, (%eax,%xmm1,2), %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_mask_i32gather_epi32: +; X64: # BB#0: +; X64-NEXT: vpgatherdd %xmm2, (%rdi,%xmm1,2), %xmm0 +; X64-NEXT: retq + %arg0 = bitcast <2 x i64> %a0 to <4 x i32> + %arg1 = bitcast i32 *%a1 to i8* + %arg2 = bitcast <2 x i64> %a2 to <4 x i32> + %arg3 = bitcast <2 x i64> %a3 to <4 x i32> + %call = call <4 x i32> @llvm.x86.avx2.gather.d.d(<4 x i32> %arg0, i8* %arg1, <4 x i32> %arg2, <4 x i32> %arg3, i8 2) + %bc = bitcast <4 x i32> %call to <2 x i64> + ret <2 x i64> %bc +} + +define <4 x i64> @test_mm256_i32gather_epi32(i32 *%a0, <4 x i64> %a1) { +; X32-LABEL: test_mm256_i32gather_epi32: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpcmpeqd %ymm2, %ymm2, %ymm2 +; X32-NEXT: vpgatherdd %ymm2, (%eax,%ymm0,2), %ymm1 +; X32-NEXT: vmovdqa %ymm1, %ymm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_i32gather_epi32: +; X64: # BB#0: +; X64-NEXT: vpcmpeqd %ymm2, %ymm2, %ymm2 +; X64-NEXT: vpgatherdd %ymm2, (%rdi,%ymm0,2), %ymm1 +; X64-NEXT: vmovdqa %ymm1, %ymm0 +; X64-NEXT: retq + %arg0 = bitcast i32 *%a0 to i8* + %arg1 = bitcast <4 x i64> %a1 to <8 x i32> + %mask = bitcast <4 x i64> <i64 -1, i64 -1, i64 -1, i64 -1> to <8 x i32> + %call = call <8 x i32> @llvm.x86.avx2.gather.d.d.256(<8 x i32> undef, i8* %arg0, <8 x i32> %arg1, <8 x i32> %mask, i8 2) + %bc = bitcast <8 x i32> %call to <4 x i64> + ret <4 x i64> %bc +} +declare <8 x i32> @llvm.x86.avx2.gather.d.d.256(<8 x i32>, i8*, <8 x i32>, <8 x i32>, i8) nounwind readonly + +define <4 x i64> @test_mm256_mask_i32gather_epi32(<4 x i64> %a0, i32 *%a1, <4 x i64> %a2, <4 x i64> %a3) { +; X32-LABEL: test_mm256_mask_i32gather_epi32: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpgatherdd %ymm2, (%eax,%ymm1,2), %ymm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_mask_i32gather_epi32: +; X64: # BB#0: +; X64-NEXT: vpgatherdd %ymm2, (%rdi,%ymm1,2), %ymm0 +; X64-NEXT: retq + %arg0 = bitcast <4 x i64> %a0 to <8 x i32> + %arg1 = bitcast i32 *%a1 to i8* + %arg2 = bitcast <4 x i64> %a2 to <8 x i32> + %arg3 = bitcast <4 x i64> %a3 to <8 x i32> + %call = call <8 x i32> @llvm.x86.avx2.gather.d.d.256(<8 x i32> %arg0, i8* %arg1, <8 x i32> %arg2, <8 x i32> %arg3, i8 2) + %bc = bitcast <8 x i32> %call to <4 x i64> + ret <4 x i64> %bc +} + +define <2 x i64> @test_mm_i32gather_epi64(i64 *%a0, <2 x i64> %a1) { +; X32-LABEL: test_mm_i32gather_epi64: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 +; X32-NEXT: vpgatherdq %xmm2, (%eax,%xmm0,2), %xmm1 +; X32-NEXT: vmovdqa %xmm1, %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_i32gather_epi64: +; X64: # BB#0: +; X64-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 +; X64-NEXT: vpgatherdq %xmm2, (%rdi,%xmm0,2), %xmm1 +; X64-NEXT: vmovdqa %xmm1, %xmm0 +; X64-NEXT: retq + %arg0 = bitcast i64 *%a0 to i8* + %arg1 = bitcast <2 x i64> %a1 to <4 x i32> + %res = call <2 x i64> @llvm.x86.avx2.gather.d.q(<2 x i64> undef, i8* %arg0, <4 x i32> %arg1, <2 x i64> <i64 -1, i64 -1>, i8 2) + ret <2 x i64> %res +} +declare <2 x i64> @llvm.x86.avx2.gather.d.q(<2 x i64>, i8*, <4 x i32>, <2 x i64>, i8) nounwind readonly + +define <2 x i64> @test_mm_mask_i32gather_epi64(<2 x i64> %a0, i64 *%a1, <2 x i64> %a2, <2 x i64> %a3) { +; X32-LABEL: test_mm_mask_i32gather_epi64: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpgatherdq %xmm2, (%eax,%xmm1,2), %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_mask_i32gather_epi64: +; X64: # BB#0: +; X64-NEXT: vpgatherdq %xmm2, (%rdi,%xmm1,2), %xmm0 +; X64-NEXT: retq + %arg1 = bitcast i64 *%a1 to i8* + %arg2 = bitcast <2 x i64> %a2 to <4 x i32> + %res = call <2 x i64> @llvm.x86.avx2.gather.d.q(<2 x i64> %a0, i8* %arg1, <4 x i32> %arg2, <2 x i64> %a3, i8 2) + ret <2 x i64> %res +} + +define <4 x i64> @test_mm256_i32gather_epi64(i64 *%a0, <2 x i64> %a1) { +; X32-LABEL: test_mm256_i32gather_epi64: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpcmpeqd %ymm2, %ymm2, %ymm2 +; X32-NEXT: vpgatherdq %ymm2, (%eax,%xmm0,2), %ymm1 +; X32-NEXT: vmovdqa %ymm1, %ymm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_i32gather_epi64: +; X64: # BB#0: +; X64-NEXT: vpcmpeqd %ymm2, %ymm2, %ymm2 +; X64-NEXT: vpgatherdq %ymm2, (%rdi,%xmm0,2), %ymm1 +; X64-NEXT: vmovdqa %ymm1, %ymm0 +; X64-NEXT: retq + %arg0 = bitcast i64 *%a0 to i8* + %arg1 = bitcast <2 x i64> %a1 to <4 x i32> + %res = call <4 x i64> @llvm.x86.avx2.gather.d.q.256(<4 x i64> undef, i8* %arg0, <4 x i32> %arg1, <4 x i64> <i64 -1, i64 -1, i64 -1, i64 -1>, i8 2) + ret <4 x i64> %res +} +declare <4 x i64> @llvm.x86.avx2.gather.d.q.256(<4 x i64>, i8*, <4 x i32>, <4 x i64>, i8) nounwind readonly + +define <4 x i64> @test_mm256_mask_i32gather_epi64(<4 x i64> %a0, i64 *%a1, <2 x i64> %a2, <4 x i64> %a3) { +; X32-LABEL: test_mm256_mask_i32gather_epi64: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpgatherdq %ymm2, (%eax,%xmm1,2), %ymm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_mask_i32gather_epi64: +; X64: # BB#0: +; X64-NEXT: vpgatherdq %ymm2, (%rdi,%xmm1,2), %ymm0 +; X64-NEXT: retq + %arg1 = bitcast i64 *%a1 to i8* + %arg2 = bitcast <2 x i64> %a2 to <4 x i32> + %res = call <4 x i64> @llvm.x86.avx2.gather.d.q.256(<4 x i64> %a0, i8* %arg1, <4 x i32> %arg2, <4 x i64> %a3, i8 2) + ret <4 x i64> %res +} + +define <2 x double> @test_mm_i32gather_pd(double *%a0, <2 x i64> %a1) { +; X32-LABEL: test_mm_i32gather_pd: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vxorpd %xmm1, %xmm1, %xmm1 +; X32-NEXT: vcmpeqpd %xmm1, %xmm1, %xmm2 +; X32-NEXT: vgatherdpd %xmm2, (%eax,%xmm0,2), %xmm1 +; X32-NEXT: vmovapd %xmm1, %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_i32gather_pd: +; X64: # BB#0: +; X64-NEXT: vxorpd %xmm1, %xmm1, %xmm1 +; X64-NEXT: vcmpeqpd %xmm1, %xmm1, %xmm2 +; X64-NEXT: vgatherdpd %xmm2, (%rdi,%xmm0,2), %xmm1 +; X64-NEXT: vmovapd %xmm1, %xmm0 +; X64-NEXT: retq + %arg0 = bitcast double *%a0 to i8* + %arg1 = bitcast <2 x i64> %a1 to <4 x i32> + %mask = call <2 x double> @llvm.x86.sse2.cmp.pd(<2 x double> zeroinitializer, <2 x double> zeroinitializer, i8 0) + %res = call <2 x double> @llvm.x86.avx2.gather.d.pd(<2 x double> undef, i8* %arg0, <4 x i32> %arg1, <2 x double> %mask, i8 2) + ret <2 x double> %res +} +declare <2 x double> @llvm.x86.avx2.gather.d.pd(<2 x double>, i8*, <4 x i32>, <2 x double>, i8) nounwind readonly + +define <2 x double> @test_mm_mask_i32gather_pd(<2 x double> %a0, double *%a1, <2 x i64> %a2, <2 x double> %a3) { +; X32-LABEL: test_mm_mask_i32gather_pd: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vgatherdpd %xmm2, (%eax,%xmm1,2), %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_mask_i32gather_pd: +; X64: # BB#0: +; X64-NEXT: vgatherdpd %xmm2, (%rdi,%xmm1,2), %xmm0 +; X64-NEXT: retq + %arg1 = bitcast double *%a1 to i8* + %arg2 = bitcast <2 x i64> %a2 to <4 x i32> + %res = call <2 x double> @llvm.x86.avx2.gather.d.pd(<2 x double> %a0, i8* %arg1, <4 x i32> %arg2, <2 x double> %a3, i8 2) + ret <2 x double> %res +} + +define <4 x double> @test_mm256_i32gather_pd(double *%a0, <2 x i64> %a1) { +; X32-LABEL: test_mm256_i32gather_pd: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vxorpd %ymm1, %ymm1, %ymm1 +; X32-NEXT: vcmpeqpd %ymm1, %ymm1, %ymm2 +; X32-NEXT: vgatherdpd %ymm2, (%eax,%xmm0,2), %ymm1 +; X32-NEXT: vmovapd %ymm1, %ymm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_i32gather_pd: +; X64: # BB#0: +; X64-NEXT: vxorpd %ymm1, %ymm1, %ymm1 +; X64-NEXT: vcmpeqpd %ymm1, %ymm1, %ymm2 +; X64-NEXT: vgatherdpd %ymm2, (%rdi,%xmm0,2), %ymm1 +; X64-NEXT: vmovapd %ymm1, %ymm0 +; X64-NEXT: retq + %arg0 = bitcast double *%a0 to i8* + %arg1 = bitcast <2 x i64> %a1 to <4 x i32> + %mask = call <4 x double> @llvm.x86.avx.cmp.pd.256(<4 x double> zeroinitializer, <4 x double> zeroinitializer, i8 0) + %res = call <4 x double> @llvm.x86.avx2.gather.d.pd.256(<4 x double> undef, i8* %arg0, <4 x i32> %arg1, <4 x double> %mask, i8 2) + ret <4 x double> %res +} +declare <4 x double> @llvm.x86.avx2.gather.d.pd.256(<4 x double>, i8*, <4 x i32>, <4 x double>, i8) nounwind readonly + +define <4 x double> @test_mm256_mask_i32gather_pd(<4 x double> %a0, double *%a1, <2 x i64> %a2, <4 x double> %a3) { +; X32-LABEL: test_mm256_mask_i32gather_pd: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vgatherdpd %ymm2, (%eax,%xmm1,2), %ymm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_mask_i32gather_pd: +; X64: # BB#0: +; X64-NEXT: vgatherdpd %ymm2, (%rdi,%xmm1,2), %ymm0 +; X64-NEXT: retq + %arg1 = bitcast double *%a1 to i8* + %arg2 = bitcast <2 x i64> %a2 to <4 x i32> + %res = call <4 x double> @llvm.x86.avx2.gather.d.pd.256(<4 x double> %a0, i8* %arg1, <4 x i32> %arg2, <4 x double> %a3, i8 2) + ret <4 x double> %res +} + +define <4 x float> @test_mm_i32gather_ps(float *%a0, <2 x i64> %a1) { +; X32-LABEL: test_mm_i32gather_ps: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vxorps %xmm1, %xmm1, %xmm1 +; X32-NEXT: vcmpeqps %xmm1, %xmm1, %xmm2 +; X32-NEXT: vgatherdps %xmm2, (%eax,%xmm0,2), %xmm1 +; X32-NEXT: vmovaps %xmm1, %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_i32gather_ps: +; X64: # BB#0: +; X64-NEXT: vxorps %xmm1, %xmm1, %xmm1 +; X64-NEXT: vcmpeqps %xmm1, %xmm1, %xmm2 +; X64-NEXT: vgatherdps %xmm2, (%rdi,%xmm0,2), %xmm1 +; X64-NEXT: vmovaps %xmm1, %xmm0 +; X64-NEXT: retq + %arg0 = bitcast float *%a0 to i8* + %arg1 = bitcast <2 x i64> %a1 to <4 x i32> + %mask = call <4 x float> @llvm.x86.sse.cmp.ps(<4 x float> zeroinitializer, <4 x float> zeroinitializer, i8 0) + %call = call <4 x float> @llvm.x86.avx2.gather.d.ps(<4 x float> undef, i8* %arg0, <4 x i32> %arg1, <4 x float> %mask, i8 2) + ret <4 x float> %call +} +declare <4 x float> @llvm.x86.avx2.gather.d.ps(<4 x float>, i8*, <4 x i32>, <4 x float>, i8) nounwind readonly + +define <4 x float> @test_mm_mask_i32gather_ps(<4 x float> %a0, float *%a1, <2 x i64> %a2, <4 x float> %a3) { +; X32-LABEL: test_mm_mask_i32gather_ps: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vgatherdps %xmm2, (%eax,%xmm1,2), %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_mask_i32gather_ps: +; X64: # BB#0: +; X64-NEXT: vgatherdps %xmm2, (%rdi,%xmm1,2), %xmm0 +; X64-NEXT: retq + %arg1 = bitcast float *%a1 to i8* + %arg2 = bitcast <2 x i64> %a2 to <4 x i32> + %call = call <4 x float> @llvm.x86.avx2.gather.d.ps(<4 x float> %a0, i8* %arg1, <4 x i32> %arg2, <4 x float> %a3, i8 2) + ret <4 x float> %call +} + +define <8 x float> @test_mm256_i32gather_ps(float *%a0, <4 x i64> %a1) { +; X32-LABEL: test_mm256_i32gather_ps: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vxorps %ymm1, %ymm1, %ymm1 +; X32-NEXT: vcmpeqps %ymm1, %ymm1, %ymm2 +; X32-NEXT: vgatherdps %ymm2, (%eax,%ymm0,2), %ymm1 +; X32-NEXT: vmovaps %ymm1, %ymm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_i32gather_ps: +; X64: # BB#0: +; X64-NEXT: vxorps %ymm1, %ymm1, %ymm1 +; X64-NEXT: vcmpeqps %ymm1, %ymm1, %ymm2 +; X64-NEXT: vgatherdps %ymm2, (%rdi,%ymm0,2), %ymm1 +; X64-NEXT: vmovaps %ymm1, %ymm0 +; X64-NEXT: retq + %arg0 = bitcast float *%a0 to i8* + %arg1 = bitcast <4 x i64> %a1 to <8 x i32> + %mask = call <8 x float> @llvm.x86.avx.cmp.ps.256(<8 x float> zeroinitializer, <8 x float> zeroinitializer, i8 0) + %call = call <8 x float> @llvm.x86.avx2.gather.d.ps.256(<8 x float> undef, i8* %arg0, <8 x i32> %arg1, <8 x float> %mask, i8 2) + ret <8 x float> %call +} +declare <8 x float> @llvm.x86.avx2.gather.d.ps.256(<8 x float>, i8*, <8 x i32>, <8 x float>, i8) nounwind readonly + +define <8 x float> @test_mm256_mask_i32gather_ps(<8 x float> %a0, float *%a1, <4 x i64> %a2, <8 x float> %a3) { +; X32-LABEL: test_mm256_mask_i32gather_ps: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vgatherdps %ymm2, (%eax,%ymm1,2), %ymm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_mask_i32gather_ps: +; X64: # BB#0: +; X64-NEXT: vgatherdps %ymm2, (%rdi,%ymm1,2), %ymm0 +; X64-NEXT: retq + %arg1 = bitcast float *%a1 to i8* + %arg2 = bitcast <4 x i64> %a2 to <8 x i32> + %call = call <8 x float> @llvm.x86.avx2.gather.d.ps.256(<8 x float> %a0, i8* %arg1, <8 x i32> %arg2, <8 x float> %a3, i8 2) + ret <8 x float> %call +} + +define <2 x i64> @test_mm_i64gather_epi32(i32 *%a0, <2 x i64> %a1) { +; X32-LABEL: test_mm_i64gather_epi32: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 +; X32-NEXT: vpgatherqd %xmm2, (%eax,%xmm0,2), %xmm1 +; X32-NEXT: vmovdqa %xmm1, %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_i64gather_epi32: +; X64: # BB#0: +; X64-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 +; X64-NEXT: vpgatherqd %xmm2, (%rdi,%xmm0,2), %xmm1 +; X64-NEXT: vmovdqa %xmm1, %xmm0 +; X64-NEXT: retq + %arg0 = bitcast i32 *%a0 to i8* + %mask = bitcast <2 x i64> <i64 -1, i64 -1> to <4 x i32> + %call = call <4 x i32> @llvm.x86.avx2.gather.q.d(<4 x i32> undef, i8* %arg0, <2 x i64> %a1, <4 x i32> %mask, i8 2) + %bc = bitcast <4 x i32> %call to <2 x i64> + ret <2 x i64> %bc +} +declare <4 x i32> @llvm.x86.avx2.gather.q.d(<4 x i32>, i8*, <2 x i64>, <4 x i32>, i8) nounwind readonly + +define <2 x i64> @test_mm_mask_i64gather_epi32(<2 x i64> %a0, i32 *%a1, <2 x i64> %a2, <2 x i64> %a3) { +; X32-LABEL: test_mm_mask_i64gather_epi32: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpgatherqd %xmm2, (%eax,%xmm1,2), %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_mask_i64gather_epi32: +; X64: # BB#0: +; X64-NEXT: vpgatherqd %xmm2, (%rdi,%xmm1,2), %xmm0 +; X64-NEXT: retq + %arg0 = bitcast <2 x i64> %a0 to <4 x i32> + %arg1 = bitcast i32 *%a1 to i8* + %arg3 = bitcast <2 x i64> %a3 to <4 x i32> + %call = call <4 x i32> @llvm.x86.avx2.gather.q.d(<4 x i32> %arg0, i8* %arg1, <2 x i64> %a2, <4 x i32> %arg3, i8 2) + %bc = bitcast <4 x i32> %call to <2 x i64> + ret <2 x i64> %bc +} + +define <2 x i64> @test_mm256_i64gather_epi32(i32 *%a0, <4 x i64> %a1) { +; X32-LABEL: test_mm256_i64gather_epi32: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 +; X32-NEXT: vpgatherqd %xmm2, (%eax,%ymm0,2), %xmm1 +; X32-NEXT: vmovdqa %xmm1, %xmm0 +; X32-NEXT: vzeroupper +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_i64gather_epi32: +; X64: # BB#0: +; X64-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 +; X64-NEXT: vpgatherqd %xmm2, (%rdi,%ymm0,2), %xmm1 +; X64-NEXT: vmovdqa %xmm1, %xmm0 +; X64-NEXT: vzeroupper +; X64-NEXT: retq + %arg0 = bitcast i32 *%a0 to i8* + %mask = bitcast <2 x i64> <i64 -1, i64 -1> to <4 x i32> + %call = call <4 x i32> @llvm.x86.avx2.gather.q.d.256(<4 x i32> undef, i8* %arg0, <4 x i64> %a1, <4 x i32> %mask, i8 2) + %bc = bitcast <4 x i32> %call to <2 x i64> + ret <2 x i64> %bc +} +declare <4 x i32> @llvm.x86.avx2.gather.q.d.256(<4 x i32>, i8*, <4 x i64>, <4 x i32>, i8) nounwind readonly + +define <2 x i64> @test_mm256_mask_i64gather_epi32(<2 x i64> %a0, i32 *%a1, <4 x i64> %a2, <2 x i64> %a3) { +; X32-LABEL: test_mm256_mask_i64gather_epi32: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpgatherqd %xmm2, (%eax,%ymm1,2), %xmm0 +; X32-NEXT: vzeroupper +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_mask_i64gather_epi32: +; X64: # BB#0: +; X64-NEXT: vpgatherqd %xmm2, (%rdi,%ymm1,2), %xmm0 +; X64-NEXT: vzeroupper +; X64-NEXT: retq + %arg0 = bitcast <2 x i64> %a0 to <4 x i32> + %arg1 = bitcast i32 *%a1 to i8* + %arg3 = bitcast <2 x i64> %a3 to <4 x i32> + %call = call <4 x i32> @llvm.x86.avx2.gather.q.d.256(<4 x i32> %arg0, i8* %arg1, <4 x i64> %a2, <4 x i32> %arg3, i8 2) + %bc = bitcast <4 x i32> %call to <2 x i64> + ret <2 x i64> %bc +} + +define <2 x i64> @test_mm_i64gather_epi64(i64 *%a0, <2 x i64> %a1) { +; X32-LABEL: test_mm_i64gather_epi64: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 +; X32-NEXT: vpgatherqq %xmm2, (%eax,%xmm0,2), %xmm1 +; X32-NEXT: vmovdqa %xmm1, %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_i64gather_epi64: +; X64: # BB#0: +; X64-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 +; X64-NEXT: vpgatherqq %xmm2, (%rdi,%xmm0,2), %xmm1 +; X64-NEXT: vmovdqa %xmm1, %xmm0 +; X64-NEXT: retq + %arg0 = bitcast i64 *%a0 to i8* + %call = call <2 x i64> @llvm.x86.avx2.gather.q.q(<2 x i64> undef, i8* %arg0, <2 x i64> %a1, <2 x i64> <i64 -1, i64 -1>, i8 2) + ret <2 x i64> %call +} +declare <2 x i64> @llvm.x86.avx2.gather.q.q(<2 x i64>, i8*, <2 x i64>, <2 x i64>, i8) nounwind readonly + +define <2 x i64> @test_mm_mask_i64gather_epi64(<2 x i64> %a0, i64 *%a1, <2 x i64> %a2, <2 x i64> %a3) { +; X32-LABEL: test_mm_mask_i64gather_epi64: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpgatherqq %xmm2, (%eax,%xmm1,2), %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_mask_i64gather_epi64: +; X64: # BB#0: +; X64-NEXT: vpgatherqq %xmm2, (%rdi,%xmm1,2), %xmm0 +; X64-NEXT: retq + %arg1 = bitcast i64 *%a1 to i8* + %call = call <2 x i64> @llvm.x86.avx2.gather.q.q(<2 x i64> %a0, i8* %arg1, <2 x i64> %a2, <2 x i64> %a3, i8 2) + ret <2 x i64> %call +} + +define <4 x i64> @test_mm256_i64gather_epi64(i64 *%a0, <4 x i64> %a1) { +; X32-LABEL: test_mm256_i64gather_epi64: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpcmpeqd %ymm2, %ymm2, %ymm2 +; X32-NEXT: vpgatherqq %ymm2, (%eax,%ymm0,2), %ymm1 +; X32-NEXT: vmovdqa %ymm1, %ymm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_i64gather_epi64: +; X64: # BB#0: +; X64-NEXT: vpcmpeqd %ymm2, %ymm2, %ymm2 +; X64-NEXT: vpgatherqq %ymm2, (%rdi,%ymm0,2), %ymm1 +; X64-NEXT: vmovdqa %ymm1, %ymm0 +; X64-NEXT: retq + %arg0 = bitcast i64 *%a0 to i8* + %call = call <4 x i64> @llvm.x86.avx2.gather.q.q.256(<4 x i64> undef, i8* %arg0, <4 x i64> %a1, <4 x i64> <i64 -1, i64 -1, i64 -1, i64 -1>, i8 2) + ret <4 x i64> %call +} +declare <4 x i64> @llvm.x86.avx2.gather.q.q.256(<4 x i64>, i8*, <4 x i64>, <4 x i64>, i8) nounwind readonly + +define <4 x i64> @test_mm256_mask_i64gather_epi64(<4 x i64> %a0, i64 *%a1, <4 x i64> %a2, <4 x i64> %a3) { +; X32-LABEL: test_mm256_mask_i64gather_epi64: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vpgatherqq %ymm2, (%eax,%ymm1,2), %ymm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_mask_i64gather_epi64: +; X64: # BB#0: +; X64-NEXT: vpgatherqq %ymm2, (%rdi,%ymm1,2), %ymm0 +; X64-NEXT: retq + %arg1 = bitcast i64 *%a1 to i8* + %call = call <4 x i64> @llvm.x86.avx2.gather.q.q.256(<4 x i64> %a0, i8* %arg1, <4 x i64> %a2, <4 x i64> %a3, i8 2) + ret <4 x i64> %call +} + +define <2 x double> @test_mm_i64gather_pd(double *%a0, <2 x i64> %a1) { +; X32-LABEL: test_mm_i64gather_pd: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vxorpd %xmm1, %xmm1, %xmm1 +; X32-NEXT: vcmpeqpd %xmm1, %xmm1, %xmm2 +; X32-NEXT: vgatherqpd %xmm2, (%eax,%xmm0,2), %xmm1 +; X32-NEXT: vmovapd %xmm1, %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_i64gather_pd: +; X64: # BB#0: +; X64-NEXT: vxorpd %xmm1, %xmm1, %xmm1 +; X64-NEXT: vcmpeqpd %xmm1, %xmm1, %xmm2 +; X64-NEXT: vgatherqpd %xmm2, (%rdi,%xmm0,2), %xmm1 +; X64-NEXT: vmovapd %xmm1, %xmm0 +; X64-NEXT: retq + %arg0 = bitcast double *%a0 to i8* + %mask = call <2 x double> @llvm.x86.sse2.cmp.pd(<2 x double> zeroinitializer, <2 x double> zeroinitializer, i8 0) + %call = call <2 x double> @llvm.x86.avx2.gather.q.pd(<2 x double> undef, i8* %arg0, <2 x i64> %a1, <2 x double> %mask, i8 2) + ret <2 x double> %call +} +declare <2 x double> @llvm.x86.avx2.gather.q.pd(<2 x double>, i8*, <2 x i64>, <2 x double>, i8) nounwind readonly + +define <2 x double> @test_mm_mask_i64gather_pd(<2 x double> %a0, double *%a1, <2 x i64> %a2, <2 x double> %a3) { +; X32-LABEL: test_mm_mask_i64gather_pd: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vgatherqpd %xmm2, (%eax,%xmm1,2), %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_mask_i64gather_pd: +; X64: # BB#0: +; X64-NEXT: vgatherqpd %xmm2, (%rdi,%xmm1,2), %xmm0 +; X64-NEXT: retq + %arg1 = bitcast double *%a1 to i8* + %call = call <2 x double> @llvm.x86.avx2.gather.q.pd(<2 x double> %a0, i8* %arg1, <2 x i64> %a2, <2 x double> %a3, i8 2) + ret <2 x double> %call +} + +define <4 x double> @test_mm256_i64gather_pd(double *%a0, <4 x i64> %a1) { +; X32-LABEL: test_mm256_i64gather_pd: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vxorpd %ymm1, %ymm1, %ymm1 +; X32-NEXT: vcmpeqpd %ymm1, %ymm1, %ymm2 +; X32-NEXT: vgatherqpd %ymm2, (%eax,%ymm0,2), %ymm1 +; X32-NEXT: vmovapd %ymm1, %ymm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_i64gather_pd: +; X64: # BB#0: +; X64-NEXT: vxorpd %ymm1, %ymm1, %ymm1 +; X64-NEXT: vcmpeqpd %ymm1, %ymm1, %ymm2 +; X64-NEXT: vgatherqpd %ymm2, (%rdi,%ymm0,2), %ymm1 +; X64-NEXT: vmovapd %ymm1, %ymm0 +; X64-NEXT: retq + %arg0 = bitcast double *%a0 to i8* + %mask = call <4 x double> @llvm.x86.avx.cmp.pd.256(<4 x double> zeroinitializer, <4 x double> zeroinitializer, i8 0) + %call = call <4 x double> @llvm.x86.avx2.gather.q.pd.256(<4 x double> undef, i8* %arg0, <4 x i64> %a1, <4 x double> %mask, i8 2) + ret <4 x double> %call +} +declare <4 x double> @llvm.x86.avx2.gather.q.pd.256(<4 x double>, i8*, <4 x i64>, <4 x double>, i8) nounwind readonly + +define <4 x double> @test_mm256_mask_i64gather_pd(<4 x double> %a0, i64 *%a1, <4 x i64> %a2, <4 x double> %a3) { +; X32-LABEL: test_mm256_mask_i64gather_pd: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vgatherqpd %ymm2, (%eax,%ymm1,2), %ymm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_mask_i64gather_pd: +; X64: # BB#0: +; X64-NEXT: vgatherqpd %ymm2, (%rdi,%ymm1,2), %ymm0 +; X64-NEXT: retq + %arg1 = bitcast i64 *%a1 to i8* + %call = call <4 x double> @llvm.x86.avx2.gather.q.pd.256(<4 x double> %a0, i8* %arg1, <4 x i64> %a2, <4 x double> %a3, i8 2) + ret <4 x double> %call +} + +define <4 x float> @test_mm_i64gather_ps(float *%a0, <2 x i64> %a1) { +; X32-LABEL: test_mm_i64gather_ps: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vxorps %xmm1, %xmm1, %xmm1 +; X32-NEXT: vcmpeqps %xmm1, %xmm1, %xmm2 +; X32-NEXT: vgatherqps %xmm2, (%eax,%xmm0,2), %xmm1 +; X32-NEXT: vmovaps %xmm1, %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_i64gather_ps: +; X64: # BB#0: +; X64-NEXT: vxorps %xmm1, %xmm1, %xmm1 +; X64-NEXT: vcmpeqps %xmm1, %xmm1, %xmm2 +; X64-NEXT: vgatherqps %xmm2, (%rdi,%xmm0,2), %xmm1 +; X64-NEXT: vmovaps %xmm1, %xmm0 +; X64-NEXT: retq + %arg0 = bitcast float *%a0 to i8* + %mask = call <4 x float> @llvm.x86.sse.cmp.ps(<4 x float> zeroinitializer, <4 x float> zeroinitializer, i8 0) + %call = call <4 x float> @llvm.x86.avx2.gather.q.ps(<4 x float> undef, i8* %arg0, <2 x i64> %a1, <4 x float> %mask, i8 2) + ret <4 x float> %call +} +declare <4 x float> @llvm.x86.avx2.gather.q.ps(<4 x float>, i8*, <2 x i64>, <4 x float>, i8) nounwind readonly + +define <4 x float> @test_mm_mask_i64gather_ps(<4 x float> %a0, float *%a1, <2 x i64> %a2, <4 x float> %a3) { +; X32-LABEL: test_mm_mask_i64gather_ps: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vgatherqps %xmm2, (%eax,%xmm1,2), %xmm0 +; X32-NEXT: retl +; +; X64-LABEL: test_mm_mask_i64gather_ps: +; X64: # BB#0: +; X64-NEXT: vgatherqps %xmm2, (%rdi,%xmm1,2), %xmm0 +; X64-NEXT: retq + %arg1 = bitcast float *%a1 to i8* + %call = call <4 x float> @llvm.x86.avx2.gather.q.ps(<4 x float> %a0, i8* %arg1, <2 x i64> %a2, <4 x float> %a3, i8 2) + ret <4 x float> %call +} + +define <4 x float> @test_mm256_i64gather_ps(float *%a0, <4 x i64> %a1) { +; X32-LABEL: test_mm256_i64gather_ps: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vxorps %xmm1, %xmm1, %xmm1 +; X32-NEXT: vcmpeqps %xmm1, %xmm1, %xmm2 +; X32-NEXT: vgatherqps %xmm2, (%eax,%ymm0,2), %xmm1 +; X32-NEXT: vmovaps %xmm1, %xmm0 +; X32-NEXT: vzeroupper +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_i64gather_ps: +; X64: # BB#0: +; X64-NEXT: vxorps %xmm1, %xmm1, %xmm1 +; X64-NEXT: vcmpeqps %xmm1, %xmm1, %xmm2 +; X64-NEXT: vgatherqps %xmm2, (%rdi,%ymm0,2), %xmm1 +; X64-NEXT: vmovaps %xmm1, %xmm0 +; X64-NEXT: vzeroupper +; X64-NEXT: retq + %arg0 = bitcast float *%a0 to i8* + %mask = call <4 x float> @llvm.x86.sse.cmp.ps(<4 x float> zeroinitializer, <4 x float> zeroinitializer, i8 0) + %call = call <4 x float> @llvm.x86.avx2.gather.q.ps.256(<4 x float> undef, i8* %arg0, <4 x i64> %a1, <4 x float> %mask, i8 2) + ret <4 x float> %call +} +declare <4 x float> @llvm.x86.avx2.gather.q.ps.256(<4 x float>, i8*, <4 x i64>, <4 x float>, i8) nounwind readonly + +define <4 x float> @test_mm256_mask_i64gather_ps(<4 x float> %a0, float *%a1, <4 x i64> %a2, <4 x float> %a3) { +; X32-LABEL: test_mm256_mask_i64gather_ps: +; X32: # BB#0: +; X32-NEXT: movl {{[0-9]+}}(%esp), %eax +; X32-NEXT: vgatherqps %xmm2, (%eax,%ymm1,2), %xmm0 +; X32-NEXT: vzeroupper +; X32-NEXT: retl +; +; X64-LABEL: test_mm256_mask_i64gather_ps: +; X64: # BB#0: +; X64-NEXT: vgatherqps %xmm2, (%rdi,%ymm1,2), %xmm0 +; X64-NEXT: vzeroupper +; X64-NEXT: retq + %arg1 = bitcast float *%a1 to i8* + %call = call <4 x float> @llvm.x86.avx2.gather.q.ps.256(<4 x float> %a0, i8* %arg1, <4 x i64> %a2, <4 x float> %a3, i8 2) + ret <4 x float> %call +} define <4 x i64> @test0_mm256_inserti128_si256(<4 x i64> %a0, <2 x i64> %a1) nounwind { ; X32-LABEL: test0_mm256_inserti128_si256: @@ -2739,3 +3366,9 @@ define <4 x i64> @test_mm256_xor_si256(<4 x i64> %a0, <4 x i64> %a1) nounwind { %res = xor <4 x i64> %a0, %a1 ret <4 x i64> %res } + +declare <4 x float> @llvm.x86.sse.cmp.ps(<4 x float>, <4 x float>, i8) nounwind readnone +declare <8 x float> @llvm.x86.avx.cmp.ps.256(<8 x float>, <8 x float>, i8) nounwind readnone + +declare <2 x double> @llvm.x86.sse2.cmp.pd(<2 x double>, <2 x double>, i8) nounwind readnone +declare <4 x double> @llvm.x86.avx.cmp.pd.256(<4 x double>, <4 x double>, i8) nounwind readnone |
