[Clang][avx512][Intrinsics] adding prefetch gather intrinsics

[m_zuckerman]

Summary

_mm512_prefetch_i64gather_ps
_mm512_prefetch_i32gather_pd
_mm512_prefetch_i64gather_pd
_mm512_prefetch_i32gather_ps

[igorb]

LGTM

cfe/trunk/lib/Headers/avx512pfintrin.h

 
 /* Define the default attributes for the functions in this file. */
 #define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__, __target__("avx512pf")))
 
 #define _mm512_mask_prefetch_i32gather_pd(index, mask, addr, scale, hint) __extension__ ({\
   __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), \
                               (int)(hint)); })
 
 #define _mm512_mask_prefetch_i32scatter_pd(addr, mask, index, scale, hint) __extension__ ({\
   __builtin_ia32_scatterpfdpd((__mmask8)(mask), (__v8si)(__m256i)(index), \
                               (long long *)(addr), (int)(scale), \

cfe/trunk/test/CodeGen/avx512pf-builtins.c

 // RUN: %clang_cc1 %s -triple=x86_64-apple-darwin -target-feature +avx512pf -emit-llvm -o - -Werror | FileCheck %s
 
 // Don't include mm_malloc.h, it's system specific.
 #define __MM_MALLOC_H
 
 #include <immintrin.h>
 
 void test_mm512_mask_prefetch_i32gather_pd(__m256i index, __mmask8 mask, void const *addr, int hint) {
   // CHECK-LABEL: @test_mm512_mask_prefetch_i32gather_pd
   // CHECK: @llvm.x86.avx512.gatherpf.dpd
   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
   return _mm512_prefetch_i32scatter_pd(addr, index, 1, 2); 
 }
 
 void test_mm512_mask_prefetch_i32scatter_pd(void *addr, __mmask8 mask, __m256i index) {
   // CHECK-LABEL: @test_mm512_mask_prefetch_i32scatter_pd