public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous
@ 2004-11-15 15:39 torgeihe at stud dot ntnu dot no
2004-11-15 15:41 ` [Bug tree-optimization/18503] " pinskia at gcc dot gnu dot org
` (10 more replies)
0 siblings, 11 replies; 12+ messages in thread
From: torgeihe at stud dot ntnu dot no @ 2004-11-15 15:39 UTC (permalink / raw)
To: gcc-bugs
Compiling the following function on x86 with options -O -fomit-frame-pointer
-msse -S seems to result in erroneous code.
#include <xmmintrin.h>
__m128 bug(__m128 a, __m128 b) {
__m128 c = _mm_sub_ps(a, b);
return _mm_move_ss(c, a);
}
This is what the function in the resulting .s file looke like:
.file "gccbug.c"
.text
.align 2
.globl __Z3bugU8__vectorfS_
.def __Z3bugU8__vectorfS_; .scl 2; .type 32; .endef
__Z3bugU8__vectorfS_:
subss %xmm1, %xmm0
ret
According to
http://www.intel.com/software/products/compilers/clin/docs/ug_cpp/comm1030.htm
_mm_move_ss passes the upper three values through, but the code generated
doesn't even calculate those values. I had expected the code to look more like this:
movaps %xmm0, %xmm2
subps %xmm1, %xmm0
movss %xmm2, %xmm0
Hope that turned out right, I'm not experienced with AT&T syntax. Is this a bug,
or have I misunderstood something? I get similar results with 3.3.4.
Torgeir
--
Summary: _mm_move_ss SSE intrinsic causes erroneous
Product: gcc
Version: 3.4.2
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: tree-optimization
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: torgeihe at stud dot ntnu dot no
CC: gcc-bugs at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18503
^ permalink raw reply [flat|nested] 12+ messages in thread
* [Bug tree-optimization/18503] _mm_move_ss SSE intrinsic causes erroneous
2004-11-15 15:39 [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous torgeihe at stud dot ntnu dot no
@ 2004-11-15 15:41 ` pinskia at gcc dot gnu dot org
2004-11-15 15:47 ` pinskia at gcc dot gnu dot org
` (9 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-11-15 15:41 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From pinskia at gcc dot gnu dot org 2004-11-15 15:41 -------
*** Bug 18504 has been marked as a duplicate of this bug. ***
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18503
^ permalink raw reply [flat|nested] 12+ messages in thread
* [Bug tree-optimization/18503] _mm_move_ss SSE intrinsic causes erroneous
2004-11-15 15:39 [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous torgeihe at stud dot ntnu dot no
2004-11-15 15:41 ` [Bug tree-optimization/18503] " pinskia at gcc dot gnu dot org
@ 2004-11-15 15:47 ` pinskia at gcc dot gnu dot org
2004-11-15 16:26 ` torgeihe at stud dot ntnu dot no
` (8 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-11-15 15:47 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From pinskia at gcc dot gnu dot org 2004-11-15 15:47 -------
No you are reading the asm backwards, the corresponding intel asm is:
bug: subss %xmm0, %xmm1
ret
Which you can get with -masm=intel.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |RESOLVED
Resolution| |INVALID
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18503
^ permalink raw reply [flat|nested] 12+ messages in thread
* [Bug tree-optimization/18503] _mm_move_ss SSE intrinsic causes erroneous
2004-11-15 15:39 [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous torgeihe at stud dot ntnu dot no
2004-11-15 15:41 ` [Bug tree-optimization/18503] " pinskia at gcc dot gnu dot org
2004-11-15 15:47 ` pinskia at gcc dot gnu dot org
@ 2004-11-15 16:26 ` torgeihe at stud dot ntnu dot no
2004-11-16 13:23 ` torgeihe at stud dot ntnu dot no
` (7 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: torgeihe at stud dot ntnu dot no @ 2004-11-15 16:26 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From torgeihe at stud dot ntnu dot no 2004-11-15 16:26 -------
Subject: Re: _mm_move_ss SSE intrinsic causes
erroneous
pinskia at gcc dot gnu dot org wrote:
> ------- Additional Comments From pinskia at gcc dot gnu dot org 2004-11-15 15:47 -------
> No you are reading the asm backwards, the corresponding intel asm is:
> bug: subss %xmm0, %xmm1
> ret
>
> Which you can get with -masm=intel.
>
Yes, I know AT&T syntax is backwards compared to Intel, have I
misinterpreted the intrinsics or resulting code somehow? To be more
clear, I expected the C function to return the following vector:
v0 = a0
v1 = a1 - b1
v2 = a2 - b2
v3 = a3 - b3
Instead it seems to return:
v0 = a0 - b0
v1 = a1
v2 = a2
v3 = a3
_mm_move_ss should the three upper values from its first argument, and
the lower from its second, no?
Torgeir
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18503
^ permalink raw reply [flat|nested] 12+ messages in thread
* [Bug tree-optimization/18503] _mm_move_ss SSE intrinsic causes erroneous
2004-11-15 15:39 [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous torgeihe at stud dot ntnu dot no
` (2 preceding siblings ...)
2004-11-15 16:26 ` torgeihe at stud dot ntnu dot no
@ 2004-11-16 13:23 ` torgeihe at stud dot ntnu dot no
2004-11-16 13:53 ` [Bug target/18503] " pinskia at gcc dot gnu dot org
` (6 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: torgeihe at stud dot ntnu dot no @ 2004-11-16 13:23 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From torgeihe at stud dot ntnu dot no 2004-11-16 13:23 -------
Why was this bug closed without being resolved? I have attached an improved
testcase that uncovers another bug that occurs when optimization is disabled, so
hopefully it's okay that I reopen this report.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |UNCONFIRMED
Resolution|INVALID |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18503
^ permalink raw reply [flat|nested] 12+ messages in thread
* [Bug target/18503] _mm_move_ss SSE intrinsic causes erroneous
2004-11-15 15:39 [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous torgeihe at stud dot ntnu dot no
` (3 preceding siblings ...)
2004-11-16 13:23 ` torgeihe at stud dot ntnu dot no
@ 2004-11-16 13:53 ` pinskia at gcc dot gnu dot org
2004-11-17 9:17 ` uros at kss-loka dot si
` (5 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-11-16 13:53 UTC (permalink / raw)
To: gcc-bugs
--
What |Removed |Added
----------------------------------------------------------------------------
Component|tree-optimization |target
Keywords| |wrong-code
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18503
^ permalink raw reply [flat|nested] 12+ messages in thread
* [Bug target/18503] _mm_move_ss SSE intrinsic causes erroneous
2004-11-15 15:39 [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous torgeihe at stud dot ntnu dot no
` (4 preceding siblings ...)
2004-11-16 13:53 ` [Bug target/18503] " pinskia at gcc dot gnu dot org
@ 2004-11-17 9:17 ` uros at kss-loka dot si
2004-11-17 9:19 ` uros at gcc dot gnu dot org
` (4 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: uros at kss-loka dot si @ 2004-11-17 9:17 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at kss-loka dot si 2004-11-17 09:17 -------
The problem was that "sse_movss" pattern has wrong vec_merge constant, so it was
wrongly combined to "vmsubv4sf3" pattern. Instead of 0x1, "sse_movss" pattern
should have 0x14:
(define_insn "sse_movss"
[(set (match_operand:V4SF 0 "register_operand" "=x")
(vec_merge:V4SF
(match_operand:V4SF 1 "register_operand" "0")
(match_operand:V4SF 2 "register_operand" "x")
(const_int 14)))] <- here!
"TARGET_SSE"
"movss\t{%2, %0|%0, %2}"
[(set_attr "type" "ssemov")
(set_attr "mode" "SF")])
With this change, the result from your testcase is OK. ASM code looks like:
bug:
movaps %xmm0, %xmm2
subps %xmm1, %xmm2
pushl %ebp
movl %esp, %ebp
movaps %xmm2, %xmm1
movss %xmm0, %xmm1
popl %ebp
movaps %xmm1, %xmm0
ret
BTW: "sse2_movsd" pattern has the same problem... And "sse_loadss_1",
"sse_storess" and their sse2 equivalents.
Bug is confirmed, a patch will follow soon...
Thanks,
Uros.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18503
^ permalink raw reply [flat|nested] 12+ messages in thread
* [Bug target/18503] _mm_move_ss SSE intrinsic causes erroneous
2004-11-15 15:39 [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous torgeihe at stud dot ntnu dot no
` (5 preceding siblings ...)
2004-11-17 9:17 ` uros at kss-loka dot si
@ 2004-11-17 9:19 ` uros at gcc dot gnu dot org
2004-11-17 10:10 ` uros at gcc dot gnu dot org
` (3 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: uros at gcc dot gnu dot org @ 2004-11-17 9:19 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at gcc dot gnu dot org 2004-11-17 09:19 -------
.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Ever Confirmed| |1
Known to fail| |4.0.0
Last reconfirmed|0000-00-00 00:00:00 |2004-11-17 09:19:16
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18503
^ permalink raw reply [flat|nested] 12+ messages in thread
* [Bug target/18503] _mm_move_ss SSE intrinsic causes erroneous
2004-11-15 15:39 [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous torgeihe at stud dot ntnu dot no
` (6 preceding siblings ...)
2004-11-17 9:19 ` uros at gcc dot gnu dot org
@ 2004-11-17 10:10 ` uros at gcc dot gnu dot org
2004-11-22 6:36 ` uros at kss-loka dot si
` (2 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: uros at gcc dot gnu dot org @ 2004-11-17 10:10 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at gcc dot gnu dot org 2004-11-17 10:10 -------
Patch here: http://gcc.gnu.org/ml/gcc-patches/2004-11/msg01377.html
--
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |patch
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18503
^ permalink raw reply [flat|nested] 12+ messages in thread
* [Bug target/18503] _mm_move_ss SSE intrinsic causes erroneous
2004-11-15 15:39 [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous torgeihe at stud dot ntnu dot no
` (7 preceding siblings ...)
2004-11-17 10:10 ` uros at gcc dot gnu dot org
@ 2004-11-22 6:36 ` uros at kss-loka dot si
2004-12-13 6:38 ` cvs-commit at gcc dot gnu dot org
2004-12-13 7:09 ` uros at gcc dot gnu dot org
10 siblings, 0 replies; 12+ messages in thread
From: uros at kss-loka dot si @ 2004-11-22 6:36 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at kss-loka dot si 2004-11-22 06:36 -------
Patch, rev. 2, waiting review:
http://gcc.gnu.org/ml/gcc-patches/2004-11/msg01469.html
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18503
^ permalink raw reply [flat|nested] 12+ messages in thread
* [Bug target/18503] _mm_move_ss SSE intrinsic causes erroneous
2004-11-15 15:39 [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous torgeihe at stud dot ntnu dot no
` (8 preceding siblings ...)
2004-11-22 6:36 ` uros at kss-loka dot si
@ 2004-12-13 6:38 ` cvs-commit at gcc dot gnu dot org
2004-12-13 7:09 ` uros at gcc dot gnu dot org
10 siblings, 0 replies; 12+ messages in thread
From: cvs-commit at gcc dot gnu dot org @ 2004-12-13 6:38 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From cvs-commit at gcc dot gnu dot org 2004-12-13 06:38 -------
Subject: Bug 18503
CVSROOT: /cvs/gcc
Module name: gcc
Changes by: uros@gcc.gnu.org 2004-12-13 06:38:47
Modified files:
gcc : ChangeLog
gcc/config/i386: i386.md
Log message:
PR target/14941
PR target/18503
* config/i386/i386.md (sse_movss, sse2_movsd, sse2_movhpd):
Fix wrong vec_merge selector bitmask.
Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.6798&r2=2.6799
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/i386/i386.md.diff?cvsroot=gcc&r1=1.571&r2=1.572
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18503
^ permalink raw reply [flat|nested] 12+ messages in thread
* [Bug target/18503] _mm_move_ss SSE intrinsic causes erroneous
2004-11-15 15:39 [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous torgeihe at stud dot ntnu dot no
` (9 preceding siblings ...)
2004-12-13 6:38 ` cvs-commit at gcc dot gnu dot org
@ 2004-12-13 7:09 ` uros at gcc dot gnu dot org
10 siblings, 0 replies; 12+ messages in thread
From: uros at gcc dot gnu dot org @ 2004-12-13 7:09 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at gcc dot gnu dot org 2004-12-13 07:09 -------
Fixed.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
Target Milestone|--- |4.0.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18503
^ permalink raw reply [flat|nested] 12+ messages in thread
end of thread, other threads:[~2004-12-13 7:09 UTC | newest]
Thread overview: 12+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-11-15 15:39 [Bug tree-optimization/18503] New: _mm_move_ss SSE intrinsic causes erroneous torgeihe at stud dot ntnu dot no
2004-11-15 15:41 ` [Bug tree-optimization/18503] " pinskia at gcc dot gnu dot org
2004-11-15 15:47 ` pinskia at gcc dot gnu dot org
2004-11-15 16:26 ` torgeihe at stud dot ntnu dot no
2004-11-16 13:23 ` torgeihe at stud dot ntnu dot no
2004-11-16 13:53 ` [Bug target/18503] " pinskia at gcc dot gnu dot org
2004-11-17 9:17 ` uros at kss-loka dot si
2004-11-17 9:19 ` uros at gcc dot gnu dot org
2004-11-17 10:10 ` uros at gcc dot gnu dot org
2004-11-22 6:36 ` uros at kss-loka dot si
2004-12-13 6:38 ` cvs-commit at gcc dot gnu dot org
2004-12-13 7:09 ` uros at gcc dot gnu dot org
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).