Many AVX512 vector operations can broadcast from a scalar memory source.
This patch enables memory broadcast for FMA operations.
gcc/
PR target/72782
* config/i386/sse.md (VF_AVX512): New.
(avx512bcst): Likewise.
(*<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name>_bcst_1):
Likewise.
(*<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name>_bcst_2):
Likewise.
(*<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name>_bcst_3):
Likewise.
gcc/testsuite/
PR target/72782
* gcc.target/i386/avx512-fma-1.h: New file.
* gcc.target/i386/avx512-fma-2.h: Likewise.
* gcc.target/i386/avx512-fma-3.h: Likewise.
* gcc.target/i386/avx512-fma-4.h: Likewise.
* gcc.target/i386/avx512-fma-5.h: Likewise.
* gcc.target/i386/avx512-fma-6.h: Likewise.
* gcc.target/i386/avx512-fma-7.h: Likewise.
* gcc.target/i386/avx512-fma-8.h: Likewise.
* gcc.target/i386/avx512f-fmadd-df-zmm-1.c: Likewise.
* gcc.target/i386/avx512f-fmadd-sf-zmm-1.c: Likewise.
* gcc.target/i386/avx512f-fmadd-sf-zmm-2.c: Likewise.
* gcc.target/i386/avx512f-fmadd-sf-zmm-3.c: Likewise.
* gcc.target/i386/avx512f-fmadd-sf-zmm-4.c: Likewise.
* gcc.target/i386/avx512f-fmadd-sf-zmm-5.c: Likewise.
* gcc.target/i386/avx512f-fmadd-sf-zmm-6.c: Likewise.
* gcc.target/i386/avx512f-fmadd-sf-zmm-7.c: Likewise.
* gcc.target/i386/avx512f-fmadd-sf-zmm-8.c: Likewise.
* gcc.target/i386/avx512vl-fmadd-sf-xmm-1.c: Likewise.
* gcc.target/i386/avx512vl-fmadd-sf-ymm-1.c: Likewise.
From-SVN: r265288
+2018-10-18 H.J. Lu <hongjiu.lu@intel.com>
+
+ PR target/72782
+ * config/i386/sse.md (VF_AVX512): New.
+ (avx512bcst): Likewise.
+ (*<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name>_bcst_1):
+ Likewise.
+ (*<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name>_bcst_2):
+ Likewise.
+ (*<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name>_bcst_3):
+ Likewise.
+
2018-10-18 Jonathan Wakely <jwakely@redhat.com>
* doc/invoke.texi (-dumpversion): Improve grammar.
(V2DI "TARGET_AVX512VL") (V2DF "TARGET_AVX512VL")])
(define_mode_iterator VI48F_256 [V8SI V8SF V4DI V4DF])
+(define_mode_iterator VF_AVX512
+ [(V4SF "TARGET_AVX512VL") (V2DF "TARGET_AVX512VL")
+ (V8SF "TARGET_AVX512VL") (V4DF "TARGET_AVX512VL")
+ V16SF V8DF])
+
+(define_mode_attr avx512bcst
+ [(V4SF "%{1to4%}") (V2DF "%{1to2%}")
+ (V8SF "%{1to8%}") (V4DF "%{1to4%}")
+ (V16SF "%{1to16%}") (V8DF "%{1to8%}")])
+
;; Mapping from float mode to required SSE level
(define_mode_attr sse
[(SF "sse") (DF "sse2")
[(set_attr "type" "ssemuladd")
(set_attr "mode" "<MODE>")])
+(define_insn "*<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name>_bcst_1"
+ [(set (match_operand:VF_AVX512 0 "register_operand" "=v,v")
+ (fma:VF_AVX512
+ (match_operand:VF_AVX512 1 "nonimmediate_operand" "0,v")
+ (match_operand:VF_AVX512 2 "nonimmediate_operand" "v,0")
+ (vec_duplicate:VF_AVX512
+ (match_operand:<ssescalarmode> 3 "memory_operand" "m,m"))))]
+ "TARGET_AVX512F && <sd_mask_mode512bit_condition>"
+ "vfmadd213<ssemodesuffix>\t{%3<avx512bcst>, %2, %0<sd_mask_op4>|%0<sd_mask_op4>, %2, %3<avx512bcst>}"
+ [(set_attr "type" "ssemuladd")
+ (set_attr "mode" "<MODE>")])
+
+(define_insn "*<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name>_bcst_2"
+ [(set (match_operand:VF_AVX512 0 "register_operand" "=v,v")
+ (fma:VF_AVX512
+ (vec_duplicate:VF_AVX512
+ (match_operand:<ssescalarmode> 1 "memory_operand" "m,m"))
+ (match_operand:VF_AVX512 2 "nonimmediate_operand" "0,v")
+ (match_operand:VF_AVX512 3 "nonimmediate_operand" "v,0")))]
+ "TARGET_AVX512F && <sd_mask_mode512bit_condition>"
+ "@
+ vfmadd132<ssemodesuffix>\t{%1<avx512bcst>, %3, %0<sd_mask_op4>|%0<sd_mask_op4>, %3, %1<avx512bcst>}
+ vfmadd231<ssemodesuffix>\t{%1<avx512bcst>, %2, %0<sd_mask_op4>|%0<sd_mask_op4>, %2, %1<avx512bcst>}"
+ [(set_attr "type" "ssemuladd")
+ (set_attr "mode" "<MODE>")])
+
+(define_insn "*<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name>_bcst_3"
+ [(set (match_operand:VF_AVX512 0 "register_operand" "=v,v")
+ (fma:VF_AVX512
+ (match_operand:VF_AVX512 1 "nonimmediate_operand" "0,v")
+ (vec_duplicate:VF_AVX512
+ (match_operand:<ssescalarmode> 2 "memory_operand" "m,m"))
+ (match_operand:VF_AVX512 3 "nonimmediate_operand" "v,0")))]
+ "TARGET_AVX512F && <sd_mask_mode512bit_condition>"
+ "@
+ vfmadd132<ssemodesuffix>\t{%2<avx512bcst>, %3, %0<sd_mask_op4>|%0<sd_mask_op4>, %3, %2<avx512bcst>}
+ vfmadd231<ssemodesuffix>\t{%2<avx512bcst>, %1, %0<sd_mask_op4>|%0<sd_mask_op4>, %1, %2<avx512bcst>}"
+ [(set_attr "type" "ssemuladd")
+ (set_attr "mode" "<MODE>")])
+
(define_insn "<avx512>_fmadd_<mode>_mask<round_name>"
[(set (match_operand:VF_AVX512VL 0 "register_operand" "=v,v")
(vec_merge:VF_AVX512VL
+2018-10-18 H.J. Lu <hongjiu.lu@intel.com>
+
+ PR target/72782
+ * gcc.target/i386/avx512-fma-1.h: New file.
+ * gcc.target/i386/avx512-fma-2.h: Likewise.
+ * gcc.target/i386/avx512-fma-3.h: Likewise.
+ * gcc.target/i386/avx512-fma-4.h: Likewise.
+ * gcc.target/i386/avx512-fma-5.h: Likewise.
+ * gcc.target/i386/avx512-fma-6.h: Likewise.
+ * gcc.target/i386/avx512-fma-7.h: Likewise.
+ * gcc.target/i386/avx512-fma-8.h: Likewise.
+ * gcc.target/i386/avx512f-fmadd-df-zmm-1.c: Likewise.
+ * gcc.target/i386/avx512f-fmadd-sf-zmm-1.c: Likewise.
+ * gcc.target/i386/avx512f-fmadd-sf-zmm-2.c: Likewise.
+ * gcc.target/i386/avx512f-fmadd-sf-zmm-3.c: Likewise.
+ * gcc.target/i386/avx512f-fmadd-sf-zmm-4.c: Likewise.
+ * gcc.target/i386/avx512f-fmadd-sf-zmm-5.c: Likewise.
+ * gcc.target/i386/avx512f-fmadd-sf-zmm-6.c: Likewise.
+ * gcc.target/i386/avx512f-fmadd-sf-zmm-7.c: Likewise.
+ * gcc.target/i386/avx512f-fmadd-sf-zmm-8.c: Likewise.
+ * gcc.target/i386/avx512vl-fmadd-sf-xmm-1.c: Likewise.
+ * gcc.target/i386/avx512vl-fmadd-sf-ymm-1.c: Likewise.
+
2018-10-18 Tobias Burnus <burnus@net-b.de>
PR fortran/87625
--- /dev/null
+#include <immintrin.h>
+
+#define PASTER2(x,y) x##y
+#define PASTER3(x,y,z) _mm##x##_##y##_##z
+#define OP(vec, op, suffix) PASTER3 (vec, op, suffix)
+#define DUP(vec, suffix, val) PASTER3 (vec, set1, suffix) (val)
+
+type
+foo (type x, type y, SCALAR *f)
+{
+ return OP (vec, op, suffix) (x, y, DUP (vec, suffix, *f));
+}
--- /dev/null
+#include <immintrin.h>
+
+#define PASTER2(x,y) x##y
+#define PASTER3(x,y,z) _mm##x##_##y##_##z
+#define TYPE(vec) PASTER2 (__m, vec)
+#define OP(vec, op, suffix) PASTER3 (vec, op, suffix)
+#define DUP(vec, suffix, val) PASTER3 (vec, set1, suffix) (val)
+
+type
+foo (type x, type y, SCALAR *f)
+{
+ return OP (vec, op, suffix) (y, x, DUP (vec, suffix, *f));
+}
--- /dev/null
+#include <immintrin.h>
+
+#define PASTER2(x,y) x##y
+#define PASTER3(x,y,z) _mm##x##_##y##_##z
+#define TYPE(vec) PASTER2 (__m, vec)
+#define OP(vec, op, suffix) PASTER3 (vec, op, suffix)
+#define DUP(vec, suffix, val) PASTER3 (vec, set1, suffix) (val)
+
+type
+foo (type x, type y, SCALAR *f)
+{
+ return OP (vec, op, suffix) (x, DUP (vec, suffix, *f), y);
+}
--- /dev/null
+#include <immintrin.h>
+
+#define PASTER2(x,y) x##y
+#define PASTER3(x,y,z) _mm##x##_##y##_##z
+#define TYPE(vec) PASTER2 (__m, vec)
+#define OP(vec, op, suffix) PASTER3 (vec, op, suffix)
+#define DUP(vec, suffix, val) PASTER3 (vec, set1, suffix) (val)
+
+type
+foo (type x, type y, SCALAR *f)
+{
+ return OP (vec, op, suffix) (y, DUP (vec, suffix, *f), x);
+}
--- /dev/null
+#include <immintrin.h>
+
+#define PASTER2(x,y) x##y
+#define PASTER3(x,y,z) _mm##x##_##y##_##z
+#define TYPE(vec) PASTER2 (__m, vec)
+#define OP(vec, op, suffix) PASTER3 (vec, op, suffix)
+#define DUP(vec, suffix, val) PASTER3 (vec, set1, suffix) (val)
+
+type
+foo (type x, type y, SCALAR *f)
+{
+ return OP (vec, op, suffix) (DUP (vec, suffix, *f), x, y);
+}
--- /dev/null
+#include <immintrin.h>
+
+#define PASTER2(x,y) x##y
+#define PASTER3(x,y,z) _mm##x##_##y##_##z
+#define TYPE(vec) PASTER2 (__m, vec)
+#define OP(vec, op, suffix) PASTER3 (vec, op, suffix)
+#define DUP(vec, suffix, val) PASTER3 (vec, set1, suffix) (val)
+
+type
+foo (type x, type y, SCALAR *f)
+{
+ return OP (vec, op, suffix) (DUP (vec, suffix, *f), y, x);
+}
--- /dev/null
+#include <immintrin.h>
+
+#define PASTER2(x,y) x##y
+#define PASTER3(x,y,z) _mm##x##_##y##_##z
+#define TYPE(vec) PASTER2 (__m, vec)
+#define OP(vec, op, suffix) PASTER3 (vec, op, suffix)
+#define DUP(vec, suffix, val) PASTER3 (vec, set1, suffix) (val)
+
+extern SCALAR bar (void);
+
+type
+foo (type x, type y)
+{
+ SCALAR f = bar ();
+ return OP (vec, op, suffix) (x, y, DUP (vec, suffix, f));
+}
--- /dev/null
+#include <immintrin.h>
+
+#define PASTER2(x,y) x##y
+#define PASTER3(x,y,z) _mm##x##_##y##_##z
+#define TYPE(vec) PASTER2 (__m, vec)
+#define OP(vec, op, suffix) PASTER3 (vec, op, suffix)
+#define DUP(vec, suffix, val) PASTER3 (vec, set1, suffix) (val)
+
+type
+foo (type x, type y, SCALAR f)
+{
+ return OP (vec, op, suffix) (y, x, DUP (vec, suffix, f));
+}
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512f -O2" } */
+/* { dg-final { scan-assembler-times "vfmadd...pd\[ \\t\]+\\(%(?:eax|rdi|edi)\\)\\\{1to\[1-8\]+\\\}, %zmm\[0-9\]+, %zmm0" 1 } } */
+/* { dg-final { scan-assembler-not "vbroadcastsd\[^\n\]*%zmm\[0-9\]+" } } */
+
+#define type __m512d
+#define vec 512
+#define op fmadd
+#define suffix pd
+#define SCALAR double
+
+#include "avx512-fma-1.h"
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512f -O2" } */
+/* { dg-final { scan-assembler-times "vfmadd...ps\[ \\t\]+\\(%(?:eax|rdi|edi)\\)\\\{1to\[1-8\]+\\\}, %zmm\[0-9\]+, %zmm0" 1 } } */
+/* { dg-final { scan-assembler-not "vbroadcastss\[^\n\]*%zmm\[0-9\]+" } } */
+
+#define type __m512
+#define vec 512
+#define op fmadd
+#define suffix ps
+#define SCALAR float
+
+#include "avx512-fma-1.h"
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512f -O2" } */
+/* { dg-final { scan-assembler-times "vfmadd...ps\[ \\t\]+\\(%(?:eax|rdi|edi)\\)\\\{1to\[1-8\]+\\\}, %zmm\[0-9\]+, %zmm0" 1 } } */
+/* { dg-final { scan-assembler-not "vbroadcastss\[^\n\]*%zmm\[0-9\]+" } } */
+
+#define type __m512
+#define vec 512
+#define op fmadd
+#define suffix ps
+#define SCALAR float
+
+#include "avx512-fma-2.h"
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512f -O2" } */
+/* { dg-final { scan-assembler-times "vfmadd...ps\[ \\t\]+\\(%(?:eax|rdi|edi)\\)\\\{1to\[1-8\]+\\\}, %zmm\[0-9\]+, %zmm0" 1 } } */
+/* { dg-final { scan-assembler-not "vbroadcastss\[^\n\]*%zmm\[0-9\]+" } } */
+
+#define type __m512
+#define vec 512
+#define op fmadd
+#define suffix ps
+#define SCALAR float
+
+#include "avx512-fma-3.h"
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512f -O2" } */
+/* { dg-final { scan-assembler-times "vfmadd...ps\[ \\t\]+\\(%(?:eax|rdi|edi)\\)\\\{1to\[1-8\]+\\\}, %zmm\[0-9\]+, %zmm0" 1 } } */
+/* { dg-final { scan-assembler-not "vbroadcastss\[^\n\]*%zmm\[0-9\]+" } } */
+
+#define type __m512
+#define vec 512
+#define op fmadd
+#define suffix ps
+#define SCALAR float
+
+#include "avx512-fma-4.h"
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512f -O2" } */
+/* { dg-final { scan-assembler-times "vfmadd...ps\[ \\t\]+\\(%(?:eax|rdi|edi)\\)\\\{1to\[1-8\]+\\\}, %zmm\[0-9\]+, %zmm0" 1 } } */
+/* { dg-final { scan-assembler-not "vbroadcastss\[^\n\]*%zmm\[0-9\]+" } } */
+
+#define type __m512
+#define vec 512
+#define op fmadd
+#define suffix ps
+#define SCALAR float
+
+#include "avx512-fma-5.h"
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512f -O2" } */
+/* { dg-final { scan-assembler-times "vfmadd...ps\[ \\t\]+\\(%(?:eax|rdi|edi)\\)\\\{1to\[1-8\]+\\\}, %zmm\[0-9\]+, %zmm0" 1 } } */
+/* { dg-final { scan-assembler-not "vbroadcastss\[^\n\]*%zmm\[0-9\]+" } } */
+
+#define type __m512
+#define vec 512
+#define op fmadd
+#define suffix ps
+#define SCALAR float
+
+#include "avx512-fma-6.h"
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mavx512f -O2" } */
+/* { dg-final { scan-assembler-times "vbroadcastss\[^\n\]*%zmm\[0-9\]+" 1 } } */
+/* { dg-final { scan-assembler-times "vfmadd...ps\[^\n\]*%zmm\[0-9\]+" 1 } } */
+
+#define type __m512
+#define vec 512
+#define op fmadd
+#define suffix ps
+#define SCALAR float
+
+#include "avx512-fma-7.h"
--- /dev/null
+/* { dg-do compile { target { ! ia32 } } } */
+/* { dg-options "-mavx512f -O2" } */
+/* { dg-final { scan-assembler-times "vbroadcastss\[^\n\]*%zmm\[0-9\]+" 1 } } */
+/* { dg-final { scan-assembler-times "vfmadd...ps\[ \\t\]+%zmm\[0-9\]+, %zmm\[0-9\]+, %zmm0" 1 } } */
+
+#define type __m512
+#define vec 512
+#define op fmadd
+#define suffix ps
+#define SCALAR float
+
+#include "avx512-fma-8.h"
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mfma -mavx512vl -O2" } */
+/* { dg-final { scan-assembler-times "vfmadd...ps\[ \\t\]+\\(%(?:eax|rdi|edi)\\)\\\{1to\[1-8\]+\\\}, %xmm\[0-9\]+, %xmm0" 1 } } */
+/* { dg-final { scan-assembler-not "vbroadcastss\[^\n\]*%xmm\[0-9\]+" } } */
+
+#define type __m128
+#define vec
+#define op fmadd
+#define suffix ps
+#define SCALAR float
+
+#include "avx512-fma-1.h"
--- /dev/null
+/* { dg-do compile } */
+/* { dg-options "-mfma -mavx512vl -O2" } */
+/* { dg-final { scan-assembler-times "vfmadd...ps\[ \\t\]+\\(%(?:eax|rdi|edi)\\)\\\{1to\[1-8\]+\\\}, %ymm\[0-9\]+, %ymm0" 1 } } */
+/* { dg-final { scan-assembler-not "vbroadcastss\[^\n\]*%ymm\[0-9\]+" } } */
+
+#define type __m256
+#define vec 256
+#define op fmadd
+#define suffix ps
+#define SCALAR float
+
+#include "avx512-fma-1.h"