This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
[PATCH] Allow XMM16-XMM31 in vpmaddubsw
- From: Jakub Jelinek <jakub at redhat dot com>
- To: Uros Bizjak <ubizjak at gmail dot com>, Kirill Yukhin <kirill dot yukhin at gmail dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Fri, 13 May 2016 19:14:50 +0200
- Subject: [PATCH] Allow XMM16-XMM31 in vpmaddubsw
- Authentication-results: sourceware.org; auth=none
- Reply-to: Jakub Jelinek <jakub at redhat dot com>
Hi!
This is either AVX2 or for EVEX AVX512BW (& AVX512VL) instruction,
thus the patch adds it as a separate alternative guarded with avx512bw
isa attribute.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
2016-05-13 Jakub Jelinek <jakub@redhat.com>
* config/i386/sse.md (avx2_pmaddubsw256, ssse3_pmaddubsw128): Add
avx512bw alternative.
* gcc.target/i386/avx512bw-vpmaddubsw-3.c: New test.
--- gcc/config/i386/sse.md.jj 2016-05-13 13:58:12.384020131 +0200
+++ gcc/config/i386/sse.md 2016-05-13 14:46:03.563465879 +0200
@@ -13933,12 +13933,12 @@ (define_insn "ssse3_ph<plusminus_mnemoni
(set_attr "mode" "DI")])
(define_insn "avx2_pmaddubsw256"
- [(set (match_operand:V16HI 0 "register_operand" "=x")
+ [(set (match_operand:V16HI 0 "register_operand" "=x,v")
(ss_plus:V16HI
(mult:V16HI
(zero_extend:V16HI
(vec_select:V16QI
- (match_operand:V32QI 1 "register_operand" "x")
+ (match_operand:V32QI 1 "register_operand" "x,v")
(parallel [(const_int 0) (const_int 2)
(const_int 4) (const_int 6)
(const_int 8) (const_int 10)
@@ -13949,7 +13949,7 @@ (define_insn "avx2_pmaddubsw256"
(const_int 28) (const_int 30)])))
(sign_extend:V16HI
(vec_select:V16QI
- (match_operand:V32QI 2 "nonimmediate_operand" "xm")
+ (match_operand:V32QI 2 "nonimmediate_operand" "xm,vm")
(parallel [(const_int 0) (const_int 2)
(const_int 4) (const_int 6)
(const_int 8) (const_int 10)
@@ -13981,9 +13981,10 @@ (define_insn "avx2_pmaddubsw256"
(const_int 29) (const_int 31)]))))))]
"TARGET_AVX2"
"vpmaddubsw\t{%2, %1, %0|%0, %1, %2}"
- [(set_attr "type" "sseiadd")
+ [(set_attr "isa" "*,avx512bw")
+ (set_attr "type" "sseiadd")
(set_attr "prefix_extra" "1")
- (set_attr "prefix" "vex")
+ (set_attr "prefix" "vex,evex")
(set_attr "mode" "OI")])
;; The correct representation for this is absolutely enormous, and
@@ -14036,19 +14037,19 @@ (define_insn "avx512bw_umulhrswv32hi3<ma
(set_attr "mode" "XI")])
(define_insn "ssse3_pmaddubsw128"
- [(set (match_operand:V8HI 0 "register_operand" "=x,x")
+ [(set (match_operand:V8HI 0 "register_operand" "=x,x,v")
(ss_plus:V8HI
(mult:V8HI
(zero_extend:V8HI
(vec_select:V8QI
- (match_operand:V16QI 1 "register_operand" "0,x")
+ (match_operand:V16QI 1 "register_operand" "0,x,v")
(parallel [(const_int 0) (const_int 2)
(const_int 4) (const_int 6)
(const_int 8) (const_int 10)
(const_int 12) (const_int 14)])))
(sign_extend:V8HI
(vec_select:V8QI
- (match_operand:V16QI 2 "vector_operand" "xBm,xm")
+ (match_operand:V16QI 2 "vector_operand" "xBm,xm,vm")
(parallel [(const_int 0) (const_int 2)
(const_int 4) (const_int 6)
(const_int 8) (const_int 10)
@@ -14069,13 +14070,14 @@ (define_insn "ssse3_pmaddubsw128"
"TARGET_SSSE3"
"@
pmaddubsw\t{%2, %0|%0, %2}
+ vpmaddubsw\t{%2, %1, %0|%0, %1, %2}
vpmaddubsw\t{%2, %1, %0|%0, %1, %2}"
- [(set_attr "isa" "noavx,avx")
+ [(set_attr "isa" "noavx,avx,avx512bw")
(set_attr "type" "sseiadd")
(set_attr "atom_unit" "simul")
- (set_attr "prefix_data16" "1,*")
+ (set_attr "prefix_data16" "1,*,*")
(set_attr "prefix_extra" "1")
- (set_attr "prefix" "orig,vex")
+ (set_attr "prefix" "orig,vex,evex")
(set_attr "mode" "TI")])
(define_insn "ssse3_pmaddubsw"
--- gcc/testsuite/gcc.target/i386/avx512bw-vpmaddubsw-3.c.jj 2016-05-13 14:55:48.714665418 +0200
+++ gcc/testsuite/gcc.target/i386/avx512bw-vpmaddubsw-3.c 2016-05-13 14:54:55.567374760 +0200
@@ -0,0 +1,30 @@
+/* { dg-do compile { target { ! ia32 } } } */
+/* { dg-options "-O2 -mavx512vl -mavx512bw" } */
+
+#include <x86intrin.h>
+
+void
+f1 (__m128i x, __m128i y)
+{
+ register __m128i a __asm ("xmm16"), b __asm ("xmm17");
+ a = x;
+ b = y;
+ asm volatile ("" : "+v" (a), "+v" (b));
+ a = _mm_maddubs_epi16 (a, b);
+ asm volatile ("" : "+v" (a));
+}
+
+/* { dg-final { scan-assembler "vpmaddubsw\[^\n\r]*xmm1\[67]\[^\n\r]*xmm1\[67]\[^\n\r]*xmm1\[67]" } } */
+
+void
+f2 (__m256i x, __m256i y)
+{
+ register __m256i a __asm ("xmm16"), b __asm ("xmm17");
+ a = x;
+ b = y;
+ asm volatile ("" : "+v" (a), "+v" (b));
+ a = _mm256_maddubs_epi16 (a, b);
+ asm volatile ("" : "+v" (a));
+}
+
+/* { dg-final { scan-assembler "vpmaddubsw\[^\n\r]*ymm1\[67]\[^\n\r]*ymm1\[67]\[^\n\r]*ymm1\[67]" } } */
Jakub