public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
@ 2005-01-18 15:44 uros at kss-loka dot si
2005-01-18 15:48 ` [Bug target/19506] [4.0 Regression] " pinskia at gcc dot gnu dot org
` (14 more replies)
0 siblings, 15 replies; 16+ messages in thread
From: uros at kss-loka dot si @ 2005-01-18 15:44 UTC (permalink / raw)
To: gcc-bugs
I think it is a good idea to open a PR for this.
There is the problem description:
http://gcc.gnu.org/ml/gcc-patches/2005-01/msg00797.html
A kind of "fix" that shows that problem is related to FP compares:
http://gcc.gnu.org/ml/gcc-patches/2005-01/msg01027.html
A comment from Richard Henderson:
http://gcc.gnu.org/ml/gcc-patches/2005-01/msg01087.html
The proposed NaN trap didn't trigger, so there is no NaNs compared:
http://gcc.gnu.org/ml/gcc-patches/2005-01/msg01096.html
This PR is not related to similar PR19496.
--
Summary: PovRay produces wrong pictures with -mfpmath=sse -ffast-
math.
Product: gcc
Version: 4.0.0
Status: UNCONFIRMED
Keywords: wrong-code
Severity: critical
Priority: P2
Component: target
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: uros at kss-loka dot si
CC: gcc-bugs at gcc dot gnu dot org
GCC build triplet: i686-pc-linux-gnu
GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
@ 2005-01-18 15:48 ` pinskia at gcc dot gnu dot org
2005-01-18 15:58 ` uros at kss-loka dot si
` (13 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2005-01-18 15:48 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From pinskia at gcc dot gnu dot org 2005-01-18 15:48 -------
I should note that -ffast-math enables -ffinite-math-only also.
--
What |Removed |Added
----------------------------------------------------------------------------
Severity|critical |normal
Summary|PovRay produces wrong |[4.0 Regression] PovRay
|pictures with -mfpmath=sse -|produces wrong pictures with
|ffast-math. |-mfpmath=sse -ffast-math.
Target Milestone|--- |4.0.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
2005-01-18 15:48 ` [Bug target/19506] [4.0 Regression] " pinskia at gcc dot gnu dot org
@ 2005-01-18 15:58 ` uros at kss-loka dot si
2005-01-18 16:06 ` uros at kss-loka dot si
` (12 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: uros at kss-loka dot si @ 2005-01-18 15:58 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at kss-loka dot si 2005-01-18 15:58 -------
This patch unfortunatelly didn't help:
http://gcc.gnu.org/ml/gcc-cvs/2005-01/msg00651.html
With the patch, tracing abyss.pov generates these warnings:
Parsing.............File: ../scenes/advanced/abyss.pov Line: 434
Warning: Degenerate CSG bounding box (not used!).
......File: ../scenes/advanced/abyss.pov Line: 642
Warning: Degenerate CSG bounding box (not used!).
...............................................File:
../scenes/advanced/abyss.pov Line: 728
Warning: Degenerate CSG bounding box (not used!).
Trying to get a smaller testcase, I have found that this problem can be somehow
fixed by following change in ix86_fp_comparison_codes():
Index: i386.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/config/i386/i386.c,v
retrieving revision 1.780
diff -u -p -r1.780 i386.c
--- i386.c 18 Jan 2005 11:08:33 -0000 1.780
+++ i386.c 18 Jan 2005 15:51:18 -0000
@@ -8446,6 +8446,8 @@ ix86_fp_comparison_codes (enum rtx_code
case LT: /* LTU - CF=1 - fails on unordered */
*first_code = UNLT;
*bypass_code = UNORDERED;
+ return;
+
break;
case LE: /* LEU - CF=1 | ZF=1 - fails on unordered */
*first_code = UNLE;
This change protects *bypass_code from being cleared back to UNKNOWN by
following code at the end of function:
...
if (!TARGET_IEEE_FP)
{
*second_code = UNKNOWN;
*bypass_code = UNKNOWN;
}
}
It looks that something is wrong with LT compares.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
2005-01-18 15:48 ` [Bug target/19506] [4.0 Regression] " pinskia at gcc dot gnu dot org
2005-01-18 15:58 ` uros at kss-loka dot si
@ 2005-01-18 16:06 ` uros at kss-loka dot si
2005-01-18 17:01 ` kcook at gcc dot gnu dot org
` (11 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: uros at kss-loka dot si @ 2005-01-18 16:06 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at kss-loka dot si 2005-01-18 16:05 -------
(In reply to comment #1)
> I should note that -ffast-math enables -ffinite-math-only also.
The same (bad) results could be achieved with -funsafe-math-optimizations.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
` (2 preceding siblings ...)
2005-01-18 16:06 ` uros at kss-loka dot si
@ 2005-01-18 17:01 ` kcook at gcc dot gnu dot org
2005-01-19 8:18 ` uros at kss-loka dot si
` (10 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: kcook at gcc dot gnu dot org @ 2005-01-18 17:01 UTC (permalink / raw)
To: gcc-bugs
--
What |Removed |Added
----------------------------------------------------------------------------
CC| |kcook at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
` (3 preceding siblings ...)
2005-01-18 17:01 ` kcook at gcc dot gnu dot org
@ 2005-01-19 8:18 ` uros at kss-loka dot si
2005-01-20 11:16 ` martin at mpa-garching dot mpg dot de
` (9 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: uros at kss-loka dot si @ 2005-01-19 8:18 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at kss-loka dot si 2005-01-19 08:17 -------
This functionality is missing after FP compares rewrite:
==i386_old.md==
...
;; We can't represent the LT test directly. Do this by swapping the operands.
(define_split
[(set (match_operand:SF 0 "fp_register_operand" "")
(if_then_else:SF (lt (match_operand:SF 1 "register_operand" "")
(match_operand:SF 2 "register_operand" ""))
(match_operand:SF 3 "register_operand" "")
(match_operand:SF 4 "register_operand" "")))
(clobber (reg:CC 17))]
"reload_completed
&& ((operands_match_p (operands[1], operands[3])
&& operands_match_p (operands[2], operands[4]))
|| (operands_match_p (operands[1], operands[4])
&& operands_match_p (operands[2], operands[3])))"
[(set (reg:CCFP 17)
(compare:CCFP (match_dup 2)
(match_dup 1)))
(set (match_dup 0)
(if_then_else:SF (ge (reg:CCFP 17) (const_int 0))
(match_dup 1)
(match_dup 2)))])
...
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
` (4 preceding siblings ...)
2005-01-19 8:18 ` uros at kss-loka dot si
@ 2005-01-20 11:16 ` martin at mpa-garching dot mpg dot de
2005-01-20 13:25 ` uros at kss-loka dot si
` (8 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: martin at mpa-garching dot mpg dot de @ 2005-01-20 11:16 UTC (permalink / raw)
To: gcc-bugs
--
What |Removed |Added
----------------------------------------------------------------------------
CC| |martin at mpa-garching dot
| |mpg dot de
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
` (5 preceding siblings ...)
2005-01-20 11:16 ` martin at mpa-garching dot mpg dot de
@ 2005-01-20 13:25 ` uros at kss-loka dot si
2005-01-20 19:12 ` rth at gcc dot gnu dot org
` (7 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: uros at kss-loka dot si @ 2005-01-20 13:25 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at kss-loka dot si 2005-01-20 13:25 -------
Adding ssemmx keyword, because this problem is triggered by -mfpmath=sse.
--
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |ssemmx
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
` (6 preceding siblings ...)
2005-01-20 13:25 ` uros at kss-loka dot si
@ 2005-01-20 19:12 ` rth at gcc dot gnu dot org
2005-01-21 0:27 ` rth at gcc dot gnu dot org
` (6 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: rth at gcc dot gnu dot org @ 2005-01-20 19:12 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From rth at gcc dot gnu dot org 2005-01-20 19:12 -------
(In reply to comment #4)
> This functionality is missing after FP compares rewrite...
No, it got moved to ix86_expand_fp_movcc.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
` (7 preceding siblings ...)
2005-01-20 19:12 ` rth at gcc dot gnu dot org
@ 2005-01-21 0:27 ` rth at gcc dot gnu dot org
2005-01-21 6:48 ` uros at kss-loka dot si
` (5 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: rth at gcc dot gnu dot org @ 2005-01-21 0:27 UTC (permalink / raw)
To: gcc-bugs
--
What |Removed |Added
----------------------------------------------------------------------------
AssignedTo|unassigned at gcc dot gnu |rth at gcc dot gnu dot org
|dot org |
Status|UNCONFIRMED |ASSIGNED
Ever Confirmed| |1
Last reconfirmed|0000-00-00 00:00:00 |2005-01-21 00:26:57
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
` (8 preceding siblings ...)
2005-01-21 0:27 ` rth at gcc dot gnu dot org
@ 2005-01-21 6:48 ` uros at kss-loka dot si
2005-01-21 12:51 ` uros at kss-loka dot si
` (4 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: uros at kss-loka dot si @ 2005-01-21 6:48 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at kss-loka dot si 2005-01-21 06:48 -------
I hope this analysis is of some help:
- the patch in comment #2 fixes the problem
- if bypass_code is cleared for (code == LT) in ix86_expand_branch() _and_
ix86_fp_comparison_fcomi_cost() the problem shows again. It should be cleared in
both places, other places doesn't affect the problem.
- TARGET_SSE_MATH part in ix86_expand_fp_movcc() can be disabled without
affecting the problem.
(following is a patch that disables failure according to comment #2, but enables
it again according to this comment):
Index: i386.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/config/i386/i386.c,v
retrieving revision 1.787
diff -u -p -r1.787 i386.c
--- i386.c 20 Jan 2005 18:34:11 -0000 1.787
+++ i386.c 21 Jan 2005 06:40:22 -0000
@@ -8470,6 +8470,8 @@ ix86_fp_comparison_codes (enum rtx_code
case LT: /* LTU - CF=1 - fails on unordered */
*first_code = UNLT;
*bypass_code = UNORDERED;
+ return;
+
break;
case LE: /* LEU - CF=1 | ZF=1 - fails on unordered */
*first_code = UNLE;
@@ -8549,6 +8551,11 @@ ix86_fp_comparison_fcomi_cost (enum rtx_
if (!TARGET_CMOVE)
return 1024;
ix86_fp_comparison_codes (code, &bypass_code, &first_code, &second_code);
+
+ // ENABLE FAILURE
+ if ((code == LT) && !TARGET_IEEE_FP)
+ bypass_code = UNKNOWN;
+
return (bypass_code != UNKNOWN || second_code != UNKNOWN) + 2;
}
@@ -8605,6 +8612,8 @@ ix86_expand_fp_compare (enum rtx_code co
*bypass_test = NULL_RTX;
ix86_fp_comparison_codes (code, &bypass_code, &first_code, &second_code);
+ // if ((code == LT) && !TARGET_IEEE_FP)
+ // bypass_code = UNKNOWN;
/* Do fcomi/sahf based test when profitable. */
if ((bypass_code == UNKNOWN || bypass_test)
@@ -8801,6 +8810,9 @@ ix86_fp_jump_nontrivial_p (enum rtx_code
if (!TARGET_CMOVE)
return true;
ix86_fp_comparison_codes (code, &bypass_code, &first_code, &second_code);
+ // if ((code == LT) && !TARGET_IEEE_FP)
+ // bypass_code = UNKNOWN;
+
return bypass_code != UNKNOWN || second_code != UNKNOWN;
}
@@ -8835,6 +8847,10 @@ ix86_expand_branch (enum rtx_code code,
ix86_fp_comparison_codes (code, &bypass_code, &first_code, &second_code);
+ //ENABLE FAILURE
+ if ((code == LT) && !TARGET_IEEE_FP)
+ bypass_code = UNKNOWN;
+
/* Check whether we will use the natural sequence with one jump. If
so, we can expand jump early. Otherwise delay expansion by
creating compound insn to not confuse optimizers. */
@@ -9767,7 +9783,7 @@ ix86_expand_fp_movcc (rtx operands[])
enum rtx_code code = GET_CODE (operands[1]);
rtx tmp, compare_op, second_test, bypass_test;
- if (TARGET_SSE_MATH && SSE_FLOAT_MODE_P (mode))
+ if (TARGET_SSE_MATH && SSE_FLOAT_MODE_P (mode) && 0)
{
rtx cmp_op0, cmp_op1, if_true, if_false;
rtx clob;
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
` (9 preceding siblings ...)
2005-01-21 6:48 ` uros at kss-loka dot si
@ 2005-01-21 12:51 ` uros at kss-loka dot si
2005-01-21 22:04 ` rth at gcc dot gnu dot org
` (3 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: uros at kss-loka dot si @ 2005-01-21 12:51 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at kss-loka dot si 2005-01-21 12:51 -------
Similar problems as described in comment #2 happens for -mfpmath=sse -mno-80387.
However, a patch at http://gcc.gnu.org/ml/gcc-patches/2005-01/msg01464.html is
needed, otherwise compilation fails.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
` (10 preceding siblings ...)
2005-01-21 12:51 ` uros at kss-loka dot si
@ 2005-01-21 22:04 ` rth at gcc dot gnu dot org
2005-01-22 11:35 ` uros at kss-loka dot si
` (2 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: rth at gcc dot gnu dot org @ 2005-01-21 22:04 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From rth at gcc dot gnu dot org 2005-01-21 22:03 -------
I've found the problem, via binary search on the povray source and then
visual inspection of the file found to be miscompiled. The mov[sd]fcc_1_sse
patterns are missing an earlyclobber. The fix isn't quite as simple as just
adding '&' to the pattern, but I should have it done today.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
` (11 preceding siblings ...)
2005-01-21 22:04 ` rth at gcc dot gnu dot org
@ 2005-01-22 11:35 ` uros at kss-loka dot si
2005-01-22 23:08 ` cvs-commit at gcc dot gnu dot org
2005-01-22 23:23 ` rth at gcc dot gnu dot org
14 siblings, 0 replies; 16+ messages in thread
From: uros at kss-loka dot si @ 2005-01-22 11:35 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at kss-loka dot si 2005-01-22 11:34 -------
Should patch to sse_comparison_operator
(http://gcc.gnu.org/ml/gcc-patches/2005-01/msg01087.html) be reverted in this
case? It didn't fix the problem, and according to comment #9 the problem is not
in this area.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
` (12 preceding siblings ...)
2005-01-22 11:35 ` uros at kss-loka dot si
@ 2005-01-22 23:08 ` cvs-commit at gcc dot gnu dot org
2005-01-22 23:23 ` rth at gcc dot gnu dot org
14 siblings, 0 replies; 16+ messages in thread
From: cvs-commit at gcc dot gnu dot org @ 2005-01-22 23:08 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From cvs-commit at gcc dot gnu dot org 2005-01-22 23:08 -------
Subject: Bug 19506
CVSROOT: /cvs/gcc
Module name: gcc
Changes by: rth@gcc.gnu.org 2005-01-22 23:07:57
Modified files:
gcc : ChangeLog
gcc/config/i386: i386.c i386.md
Log message:
PR target/19506
* config/i386/i386.md (movsfcc_1_sse_max): Use nonimmediate_operand
in both compare operands.
(movdfcc_1_sse_max): Likewise.
(movsfcc_1_sse): Likewise. Add earlyclobber for scratch.
(movdfcc_1_sse): Likewise.
* config/i386/i386.c (ix86_split_sse_movcc): Emit copies into the
scratch register as needed.
Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.7238&r2=2.7239
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/i386/i386.c.diff?cvsroot=gcc&r1=1.788&r2=1.789
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/i386/i386.md.diff?cvsroot=gcc&r1=1.609&r2=1.610
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Bug target/19506] [4.0 Regression] PovRay produces wrong pictures with -mfpmath=sse -ffast-math.
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
` (13 preceding siblings ...)
2005-01-22 23:08 ` cvs-commit at gcc dot gnu dot org
@ 2005-01-22 23:23 ` rth at gcc dot gnu dot org
14 siblings, 0 replies; 16+ messages in thread
From: rth at gcc dot gnu dot org @ 2005-01-22 23:23 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From rth at gcc dot gnu dot org 2005-01-22 23:23 -------
(In reply to comment #10)
> Should patch to sse_comparison_operator [...] be reverted in this case?
Nah. As I said in that message, allowing these operators in this manner
doesn't actually give the register allocator any more freedom, and therefore
doesn't do what was hoped. See e.g. alpha's movsicc_internal for an example
of how the comparisons would have to be reworked in order to give that freedom.
Anyway, fixed.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|ASSIGNED |RESOLVED
Resolution| |FIXED
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19506
^ permalink raw reply [flat|nested] 16+ messages in thread
end of thread, other threads:[~2005-01-22 23:23 UTC | newest]
Thread overview: 16+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2005-01-18 15:44 [Bug target/19506] New: PovRay produces wrong pictures with -mfpmath=sse -ffast-math uros at kss-loka dot si
2005-01-18 15:48 ` [Bug target/19506] [4.0 Regression] " pinskia at gcc dot gnu dot org
2005-01-18 15:58 ` uros at kss-loka dot si
2005-01-18 16:06 ` uros at kss-loka dot si
2005-01-18 17:01 ` kcook at gcc dot gnu dot org
2005-01-19 8:18 ` uros at kss-loka dot si
2005-01-20 11:16 ` martin at mpa-garching dot mpg dot de
2005-01-20 13:25 ` uros at kss-loka dot si
2005-01-20 19:12 ` rth at gcc dot gnu dot org
2005-01-21 0:27 ` rth at gcc dot gnu dot org
2005-01-21 6:48 ` uros at kss-loka dot si
2005-01-21 12:51 ` uros at kss-loka dot si
2005-01-21 22:04 ` rth at gcc dot gnu dot org
2005-01-22 11:35 ` uros at kss-loka dot si
2005-01-22 23:08 ` cvs-commit at gcc dot gnu dot org
2005-01-22 23:23 ` rth 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).