Re: [PATCH] D21322: [Clang][avx512][Intrinsics] adding prefetch gather intrinsics
This revision was automatically updated to reflect the committed changes. Closed by commit rL272667: [Clang][avx512][Intrinsics] adding prefetch gather intrinsics (authored by mzuckerm). Changed prior to commit: http://reviews.llvm.org/D21322?vs=60663=60678#toc Repository: rL LLVM http://reviews.llvm.org/D21322 Files: cfe/trunk/lib/Headers/avx512pfintrin.h cfe/trunk/test/CodeGen/avx512pf-builtins.c Index: cfe/trunk/test/CodeGen/avx512pf-builtins.c === --- cfe/trunk/test/CodeGen/avx512pf-builtins.c +++ cfe/trunk/test/CodeGen/avx512pf-builtins.c @@ -11,24 +11,48 @@ return _mm512_mask_prefetch_i32gather_pd(index, mask, addr, 2, 1); } +void test_mm512_prefetch_i32gather_pd(__m256i index, void const *addr, int hint) { + // CHECK-LABEL: @test_mm512_prefetch_i32gather_pd + // CHECK: @llvm.x86.avx512.gatherpf.dpd + return _mm512_prefetch_i32gather_pd(index, addr, 2, 1); +} + void test_mm512_mask_prefetch_i32gather_ps(__m512i index, __mmask16 mask, void const *addr, int hint) { // CHECK-LABEL: @test_mm512_mask_prefetch_i32gather_ps // CHECK: @llvm.x86.avx512.gatherpf.dps return _mm512_mask_prefetch_i32gather_ps(index, mask, addr, 2, 1); } +void test_mm512_prefetch_i32gather_ps(__m512i index, void const *addr, int hint) { + // CHECK-LABEL: @test_mm512_prefetch_i32gather_ps + // CHECK: @llvm.x86.avx512.gatherpf.dps + return _mm512_prefetch_i32gather_ps(index, addr, 2, 1); +} + void test_mm512_mask_prefetch_i64gather_pd(__m512i index, __mmask8 mask, void const *addr, int hint) { // CHECK-LABEL: @test_mm512_mask_prefetch_i64gather_pd // CHECK: @llvm.x86.avx512.gatherpf.qpd return _mm512_mask_prefetch_i64gather_pd(index, mask, addr, 2, 1); } +void test_mm512_prefetch_i64gather_pd(__m512i index, void const *addr, int hint) { + // CHECK-LABEL: @test_mm512_prefetch_i64gather_pd + // CHECK: @llvm.x86.avx512.gatherpf.qpd + return _mm512_prefetch_i64gather_pd(index, addr, 2, 1); +} + void test_mm512_mask_prefetch_i64gather_ps(__m512i index, __mmask8 mask, void const *addr, int hint) { // CHECK-LABEL: @test_mm512_mask_prefetch_i64gather_ps // CHECK: @llvm.x86.avx512.gatherpf.qps return _mm512_mask_prefetch_i64gather_ps(index, mask, addr, 2, 1); } +void test_mm512_prefetch_i64gather_ps(__m512i index, void const *addr, int hint) { + // CHECK-LABEL: @test_mm512_prefetch_i64gather_ps + // CHECK: @llvm.x86.avx512.gatherpf.qps + return _mm512_prefetch_i64gather_ps(index, addr, 2, 1); +} + void test_mm512_prefetch_i32scatter_pd(void *addr, __m256i index) { // CHECK-LABEL: @test_mm512_prefetch_i32scatter_pd // CHECK: @llvm.x86.avx512.scatterpf.dpd.512 Index: cfe/trunk/lib/Headers/avx512pfintrin.h === --- cfe/trunk/lib/Headers/avx512pfintrin.h +++ cfe/trunk/lib/Headers/avx512pfintrin.h @@ -35,21 +35,40 @@ __builtin_ia32_gatherpfdpd((__mmask8)(mask), (__v8si)(__m256i)(index), \ (long long const *)(addr), (int)(scale), \ (int)(hint)); }) + +#define _mm512_prefetch_i32gather_pd(index, addr, scale, hint) __extension__ ({\ + __builtin_ia32_gatherpfdpd((__mmask8) -1, (__v8si)(__m256i)(index), \ + (long long const *)(addr), (int)(scale), \ + (int)(hint)); }) #define _mm512_mask_prefetch_i32gather_ps(index, mask, addr, scale, hint) ({\ __builtin_ia32_gatherpfdps((__mmask16)(mask), \ (__v16si)(__m512i)(index), (int const *)(addr), \ (int)(scale), (int)(hint)); }) +#define _mm512_prefetch_i32gather_ps(index, addr, scale, hint) ({\ + __builtin_ia32_gatherpfdps((__mmask16) -1, \ + (__v16si)(__m512i)(index), (int const *)(addr), \ + (int)(scale), (int)(hint)); }) + #define _mm512_mask_prefetch_i64gather_pd(index, mask, addr, scale, hint) __extension__ ({\ __builtin_ia32_gatherpfqpd((__mmask8)(mask), (__v8di)(__m512i)(index), \ (long long const *)(addr), (int)(scale), \ (int)(hint)); }) +#define _mm512_prefetch_i64gather_pd(index, addr, scale, hint) __extension__ ({\ + __builtin_ia32_gatherpfqpd((__mmask8) -1, (__v8di)(__m512i)(index), \ + (long long const *)(addr), (int)(scale), \ + (int)(hint)); }) + #define _mm512_mask_prefetch_i64gather_ps(index, mask, addr, scale, hint) ({\ __builtin_ia32_gatherpfqps((__mmask8)(mask), (__v8di)(__m512i)(index), \ (int const *)(addr), (int)(scale), (int)(hint)); }) +#define _mm512_prefetch_i64gather_ps(index, addr, scale, hint) ({\ + __builtin_ia32_gatherpfqps((__mmask8) -1, (__v8di)(__m512i)(index), \ + (int const *)(addr), (int)(scale),
Re: [PATCH] D21322: [Clang][avx512][Intrinsics] adding prefetch gather intrinsics
igorb accepted this revision. igorb added a comment. This revision is now accepted and ready to land. LGTM http://reviews.llvm.org/D21322 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[PATCH] D21322: [Clang][avx512][Intrinsics] adding prefetch gather intrinsics
m_zuckerman created this revision. m_zuckerman added reviewers: AsafBadouh, igorb, delena. m_zuckerman added a subscriber: cfe-commits. _mm512_prefetch_i64gather_ps _mm512_prefetch_i32gather_pd _mm512_prefetch_i64gather_pd _mm512_prefetch_i32gather_ps http://reviews.llvm.org/D21322 Files: lib/Headers/avx512pfintrin.h test/CodeGen/avx512pf-builtins.c Index: lib/Headers/avx512pfintrin.h === --- lib/Headers/avx512pfintrin.h +++ lib/Headers/avx512pfintrin.h @@ -35,21 +35,40 @@ __builtin_ia32_gatherpfdpd((__mmask8)(mask), (__v8si)(__m256i)(index), \ (long long const *)(addr), (int)(scale), \ (int)(hint)); }) + +#define _mm512_prefetch_i32gather_pd(index, addr, scale, hint) __extension__ ({\ + __builtin_ia32_gatherpfdpd((__mmask8) -1, (__v8si)(__m256i)(index), \ + (long long const *)(addr), (int)(scale), \ + (int)(hint)); }) #define _mm512_mask_prefetch_i32gather_ps(index, mask, addr, scale, hint) ({\ __builtin_ia32_gatherpfdps((__mmask16)(mask), \ (__v16si)(__m512i)(index), (int const *)(addr), \ (int)(scale), (int)(hint)); }) +#define _mm512_prefetch_i32gather_ps(index, addr, scale, hint) ({\ + __builtin_ia32_gatherpfdps((__mmask16) -1, \ + (__v16si)(__m512i)(index), (int const *)(addr), \ + (int)(scale), (int)(hint)); }) + #define _mm512_mask_prefetch_i64gather_pd(index, mask, addr, scale, hint) __extension__ ({\ __builtin_ia32_gatherpfqpd((__mmask8)(mask), (__v8di)(__m512i)(index), \ (long long const *)(addr), (int)(scale), \ (int)(hint)); }) +#define _mm512_prefetch_i64gather_pd(index, addr, scale, hint) __extension__ ({\ + __builtin_ia32_gatherpfqpd((__mmask8) -1, (__v8di)(__m512i)(index), \ + (long long const *)(addr), (int)(scale), \ + (int)(hint)); }) + #define _mm512_mask_prefetch_i64gather_ps(index, mask, addr, scale, hint) ({\ __builtin_ia32_gatherpfqps((__mmask8)(mask), (__v8di)(__m512i)(index), \ (int const *)(addr), (int)(scale), (int)(hint)); }) +#define _mm512_prefetch_i64gather_ps(index, addr, scale, hint) ({\ + __builtin_ia32_gatherpfqps((__mmask8) -1, (__v8di)(__m512i)(index), \ + (int const *)(addr), (int)(scale), (int)(hint)); }) + #define _mm512_prefetch_i32scatter_pd(addr, index, scale, hint) __extension__ ({\ __builtin_ia32_scatterpfdpd((__mmask8)-1, (__v8si)(__m256i)(index), \ (long long *)(addr), (int)(scale), \ Index: test/CodeGen/avx512pf-builtins.c === --- test/CodeGen/avx512pf-builtins.c +++ test/CodeGen/avx512pf-builtins.c @@ -11,24 +11,48 @@ return _mm512_mask_prefetch_i32gather_pd(index, mask, addr, 2, 1); } +void test_mm512_prefetch_i32gather_pd(__m256i index, void const *addr, int hint) { + // CHECK-LABEL: @test_mm512_prefetch_i32gather_pd + // CHECK: @llvm.x86.avx512.gatherpf.dpd + return _mm512_prefetch_i32gather_pd(index, addr, 2, 1); +} + void test_mm512_mask_prefetch_i32gather_ps(__m512i index, __mmask16 mask, void const *addr, int hint) { // CHECK-LABEL: @test_mm512_mask_prefetch_i32gather_ps // CHECK: @llvm.x86.avx512.gatherpf.dps return _mm512_mask_prefetch_i32gather_ps(index, mask, addr, 2, 1); } +void test_mm512_prefetch_i32gather_ps(__m512i index, void const *addr, int hint) { + // CHECK-LABEL: @test_mm512_prefetch_i32gather_ps + // CHECK: @llvm.x86.avx512.gatherpf.dps + return _mm512_prefetch_i32gather_ps(index, addr, 2, 1); +} + void test_mm512_mask_prefetch_i64gather_pd(__m512i index, __mmask8 mask, void const *addr, int hint) { // CHECK-LABEL: @test_mm512_mask_prefetch_i64gather_pd // CHECK: @llvm.x86.avx512.gatherpf.qpd return _mm512_mask_prefetch_i64gather_pd(index, mask, addr, 2, 1); } +void test_mm512_prefetch_i64gather_pd(__m512i index, void const *addr, int hint) { + // CHECK-LABEL: @test_mm512_prefetch_i64gather_pd + // CHECK: @llvm.x86.avx512.gatherpf.qpd + return _mm512_prefetch_i64gather_pd(index, addr, 2, 1); +} + void test_mm512_mask_prefetch_i64gather_ps(__m512i index, __mmask8 mask, void const *addr, int hint) { // CHECK-LABEL: @test_mm512_mask_prefetch_i64gather_ps // CHECK: @llvm.x86.avx512.gatherpf.qps return _mm512_mask_prefetch_i64gather_ps(index, mask, addr, 2, 1); } +void test_mm512_prefetch_i64gather_ps(__m512i index, void const *addr, int hint) { + // CHECK-LABEL: @test_mm512_prefetch_i64gather_ps + // CHECK: @llvm.x86.avx512.gatherpf.qps + return