sse.md (*sse_concatv4sf): Renamed to ...

gcc/

2008-05-12  H.J. Lu  <hongjiu.lu@intel.com>

	* config/i386/sse.md (*sse_concatv4sf): Renamed to ...
	(*vec_concatv4sf_sse): This.
	(*sse2_concatv2si): Renamed to ...
	(*vec_concatv2si_sse2): This.
	(*sse1_concatv2si): Renamed to ...
	(*vec_concatv2si_sse): This.
	(*vec_concatv2di_rex): Renamed to ...
	(*vec_concatv2di_rex64): This.
	(*vec_concatv2si_sse4_1): New.
	(*vec_concatv2di_rex64_sse4_1): Likewise.

gcc/testsuite

2008-05-12  H.J. Lu  <hongjiu.lu@intel.com>

	* gcc.target/i386/sse2-set-epi32-1.c: New.
	* gcc.target/i386/sse2-set-epi64x-1.c: Likewise.
	* gcc.target/i386/sse4_1-set-epi32-1.c: Likewise.
	* gcc.target/i386/sse4_1-set-epi64x-1.c: Likewise.

From-SVN: r135229
This commit is contained in:
H.J. Lu 2008-05-12 20:29:37 +00:00 committed by H.J. Lu
parent 162bfc7efa
commit fcc9fe1e6a
7 changed files with 215 additions and 5 deletions

View File

@ -1,3 +1,16 @@
2008-05-12 H.J. Lu <hongjiu.lu@intel.com>
* config/i386/sse.md (*sse_concatv4sf): Renamed to ...
(*vec_concatv4sf_sse): This.
(*sse2_concatv2si): Renamed to ...
(*vec_concatv2si_sse2): This.
(*sse1_concatv2si): Renamed to ...
(*vec_concatv2si_sse): This.
(*vec_concatv2di_rex): Renamed to ...
(*vec_concatv2di_rex64): This.
(*vec_concatv2si_sse4_1): New.
(*vec_concatv2di_rex64_sse4_1): Likewise.
2008-05-12 Uros Bizjak <ubizjak@gmail.com>
PR rtl-optimization/36111

View File

@ -2253,7 +2253,7 @@
;; ??? In theory we can match memory for the MMX alternative, but allowing
;; nonimmediate_operand for operand 2 and *not* allowing memory for the SSE
;; alternatives pretty much forces the MMX alternative to be chosen.
(define_insn "*sse_concatv2sf"
(define_insn "*vec_concatv2sf_sse"
[(set (match_operand:V2SF 0 "register_operand" "=x,x,*y,*y")
(vec_concat:V2SF
(match_operand:SF 1 "nonimmediate_operand" " 0,m, 0, m")
@ -2267,7 +2267,7 @@
[(set_attr "type" "sselog,ssemov,mmxcvt,mmxmov")
(set_attr "mode" "V4SF,SF,DI,DI")])
(define_insn "*sse_concatv4sf"
(define_insn "*vec_concatv4sf_sse"
[(set (match_operand:V4SF 0 "register_operand" "=x,x")
(vec_concat:V4SF
(match_operand:V2SF 1 "register_operand" " 0,0")
@ -4726,10 +4726,22 @@
[(set_attr "type" "sselog1,ssemov")
(set_attr "mode" "TI,V4SF")])
(define_insn "*vec_concatv2si_sse4_1"
[(set (match_operand:V2SI 0 "register_operand" "=x,x")
(vec_concat:V2SI
(match_operand:SI 1 "nonimmediate_operand" "0,rm")
(match_operand:SI 2 "nonimmediate_operand" "rm,0")))]
"TARGET_SSE4_1"
"@
pinsrd\t{$0x1, %2, %0|%0, %2, 0x1}
pinsrd\t{$0x0, %2, %0|%0, %2, 0x0}"
[(set_attr "type" "sselog")
(set_attr "mode" "TI")])
;; ??? In theory we can match memory for the MMX alternative, but allowing
;; nonimmediate_operand for operand 2 and *not* allowing memory for the SSE
;; alternatives pretty much forces the MMX alternative to be chosen.
(define_insn "*sse2_concatv2si"
(define_insn "*vec_concatv2si_sse2"
[(set (match_operand:V2SI 0 "register_operand" "=Y2, Y2,*y,*y")
(vec_concat:V2SI
(match_operand:SI 1 "nonimmediate_operand" " 0 ,rm , 0,rm")
@ -4743,7 +4755,7 @@
[(set_attr "type" "sselog,ssemov,mmxcvt,mmxmov")
(set_attr "mode" "TI,TI,DI,DI")])
(define_insn "*sse1_concatv2si"
(define_insn "*vec_concatv2si_sse"
[(set (match_operand:V2SI 0 "register_operand" "=x,x,*y,*y")
(vec_concat:V2SI
(match_operand:SI 1 "nonimmediate_operand" " 0,m, 0,*rm")
@ -4770,6 +4782,18 @@
[(set_attr "type" "sselog,ssemov,ssemov")
(set_attr "mode" "TI,V4SF,V2SF")])
(define_insn "*vec_concatv2di_rex64_sse4_1"
[(set (match_operand:V2DI 0 "register_operand" "=x,x")
(vec_concat:V2DI
(match_operand:DI 1 "nonimmediate_operand" "0,rm")
(match_operand:DI 2 "nonimmediate_operand" "rm,0")))]
"TARGET_64BIT && TARGET_SSE4_1"
"@
pinsrq\t{$0x1, %2, %0|%0, %2, 0x1}
pinsrq\t{$0x0, %2, %0|%0, %2, 0x0}"
[(set_attr "type" "sselog")
(set_attr "mode" "TI")])
(define_insn "vec_concatv2di"
[(set (match_operand:V2DI 0 "register_operand" "=Y2,?Y2,Y2,x,x,x")
(vec_concat:V2DI
@ -4786,7 +4810,7 @@
[(set_attr "type" "ssemov,ssemov,sselog,ssemov,ssemov,ssemov")
(set_attr "mode" "TI,TI,TI,V4SF,V2SF,V2SF")])
(define_insn "*vec_concatv2di_rex"
(define_insn "*vec_concatv2di_rex64"
[(set (match_operand:V2DI 0 "register_operand" "=Y2,Yi,!Y2,Y2,x,x,x")
(vec_concat:V2DI
(match_operand:DI 1 "nonimmediate_operand" " m,r ,*y ,0 ,0,0,m")

View File

@ -1,3 +1,10 @@
2008-05-12 H.J. Lu <hongjiu.lu@intel.com>
* gcc.target/i386/sse2-set-epi32-1.c: New.
* gcc.target/i386/sse2-set-epi64x-1.c: Likewise.
* gcc.target/i386/sse4_1-set-epi32-1.c: Likewise.
* gcc.target/i386/sse4_1-set-epi64x-1.c: Likewise.
2008-05-12 Uros Bizjak <ubizjak@gmail.com>
PR rtl-optimization/36111

View File

@ -0,0 +1,41 @@
/* { dg-do run } */
/* { dg-options "-O2 -msse2" } */
#include "sse2-check.h"
#ifdef DEBUG
#include <stdio.h>
#endif
#include <emmintrin.h>
static void
__attribute__((noinline))
test (unsigned int *v)
{
union
{
__m128i x;
unsigned int i[4];
} u;
unsigned int i;
u.x = _mm_set_epi32 (v[3], v[2], v[1], v[0]);
for (i = 0; i < sizeof (v) / sizeof (v[0]); i++)
if (v[i] != u.i[i])
{
#ifdef DEBUG
printf ("%i: 0x%x != 0x%x\n", i, v[i], u.i[i]);
#endif
abort ();
}
}
static void
sse2_test (void)
{
unsigned int v[4]
= { 0x7B5B5465, 0x73745665, 0x63746F72, 0x5D53475D };
test (v);
}

View File

@ -0,0 +1,41 @@
/* { dg-do run } */
/* { dg-options "-O2 -msse2" } */
#include "sse2-check.h"
#ifdef DEBUG
#include <stdio.h>
#endif
#include <emmintrin.h>
static void
__attribute__((noinline))
test (unsigned long long *v)
{
union
{
__m128i x;
unsigned long long i[2];
} u;
unsigned int i;
u.x = _mm_set_epi64x (v[1], v[0]);
for (i = 0; i < sizeof (v) / sizeof (v[0]); i++)
if (v[i] != u.i[i])
{
#ifdef DEBUG
printf ("%i: 0x%llx != 0x%llx\n", i, v[i], u.i[i]);
#endif
abort ();
}
}
static void
sse2_test (void)
{
unsigned long long v[2]
= { 0x7B5B546573745665LL, 0x63746F725D53475DLL };
test (v);
}

View File

@ -0,0 +1,42 @@
/* { dg-do run } */
/* { dg-require-effective-target sse4 } */
/* { dg-options "-O2 -msse4.1" } */
#include "sse4_1-check.h"
#ifdef DEBUG
#include <stdio.h>
#endif
#include <emmintrin.h>
static void
__attribute__((noinline))
test (unsigned int *v)
{
union
{
__m128i x;
unsigned int i[4];
} u;
unsigned int i;
u.x = _mm_set_epi32 (v[3], v[2], v[1], v[0]);
for (i = 0; i < sizeof (v) / sizeof (v[0]); i++)
if (v[i] != u.i[i])
{
#ifdef DEBUG
printf ("%i: 0x%x != 0x%x\n", i, v[i], u.i[i]);
#endif
abort ();
}
}
static void
sse4_1_test (void)
{
unsigned int v[4]
= { 0x7B5B5465, 0x73745665, 0x63746F72, 0x5D53475D };
test (v);
}

View File

@ -0,0 +1,42 @@
/* { dg-do run } */
/* { dg-require-effective-target sse4 } */
/* { dg-options "-O2 -msse4.1" } */
#include "sse4_1-check.h"
#ifdef DEBUG
#include <stdio.h>
#endif
#include <emmintrin.h>
static void
__attribute__((noinline))
test (unsigned long long *v)
{
union
{
__m128i x;
unsigned long long i[2];
} u;
unsigned int i;
u.x = _mm_set_epi64x (v[1], v[0]);
for (i = 0; i < sizeof (v) / sizeof (v[0]); i++)
if (v[i] != u.i[i])
{
#ifdef DEBUG
printf ("%i: 0x%llx != 0x%llx\n", i, v[i], u.i[i]);
#endif
abort ();
}
}
static void
sse4_1_test (void)
{
unsigned long long v[2]
= { 0x7B5B546573745665LL, 0x63746F725D53475DLL };
test (v);
}