--- Comment #7 from rguenth at gcc dot gnu dot org 2009-01-31 10:25 ---
Fixed.
--
rguenth at gcc dot gnu dot org changed:
What|Removed |Added
Status|ASSIGNED
--- Comment #6 from rguenth at gcc dot gnu dot org 2009-01-31 10:25 ---
Subject: Bug 38937
Author: rguenth
Date: Sat Jan 31 10:25:41 2009
New Revision: 143819
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=143819
Log:
2009-01-30 Richard Guenther rguent...@suse.de
PR
--- Comment #2 from bonzini at gnu dot org 2009-01-31 14:19 ---
Unfortunately, I do not see any reason why the two should have different speed
(which means there's no way to teach GCC the former is better).
I think a WONTFIX is the only possibility. CCing a release manager.
--
--- Comment #58 from bonzini at gnu dot org 2009-01-31 14:26 ---
Btw, on x86_64 leslie3d performance is now above that from before r126326.
Changing to 4.3 only. A separate bug (likely not a regression) should be
opened for the testcase of comment #47.
Might even be WONTFIX for 4.3.
--- Comment #3 from rguenth at gcc dot gnu dot org 2009-01-31 14:33 ---
I would say this needs a much more detailed pipeline description. (btw, what
is diffed against what? i.e. which variant is faster? ;))
As we have %r9 and %r9d access I would say this may be some artifacts in
HW
--- Comment #2 from bonzini at gnu dot org 2009-01-31 14:33 ---
??? Andrew, there's 11 vs. 12 instructions.
--
bonzini at gnu dot org changed:
What|Removed |Added
Thread model: posix
gcc version 4.4.0 20090131 (experimental) [trunk revision 143817] (GCC)
When I updated from 143454 / 143680 (broken?) to 143817 (builds OK) I
did not notice any updates to Java that would fix this so I am assuming
that the additional ./configure options caused the breakage
--- Comment #59 from rguenth at gcc dot gnu dot org 2009-01-31 14:35
---
This is btw very likely fixed on the alias-improvements branch. SPEC testing
that for PPC would be much appreciated.
OTOH I agree with Paolo, so ... FIXED for 4.4, WONTFIX for 4.3.
--
rguenth at gcc dot gnu
--- Comment #5 from bonzini at gnu dot org 2009-01-31 14:36 ---
A pass like the one Google's guy posted to lower register pressure would help
here.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38434
--- Comment #10 from bonzini at gnu dot org 2009-01-31 14:38 ---
IMHO we should either enable flag_profile_correction by default or
remove this error again.
Agreed.
--
bonzini at gnu dot org changed:
What|Removed |Added
--- Comment #2 from rob1weld at aol dot com 2009-01-31 15:17 ---
Occurs in trunk revision 143660:
gmake[4]: Entering directory
`/usr/share/src/gcc_build/i386-pc-solaris2.11/boehm-gc'
GC_push_all_stacks: sp not set!
/bin/sh: line 9: 28147: Abort(coredump)
FAIL: gctest
There is a
--- Comment #9 from bonzini at gnu dot org 2009-01-31 15:20 ---
PING?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27078
--- Comment #1 from tydeman at tybor dot com 2009-01-31 15:29 ---
Problem still exists in gcc 4.3.2-7
--
tydeman at tybor dot com changed:
What|Removed |Added
--- Comment #8 from bonzini at gnu dot org 2009-01-31 15:32 ---
Can you check comment #6?
--
bonzini at gnu dot org changed:
What|Removed |Added
Status|NEW
--- Comment #4 from bonzini at gnu dot org 2009-01-31 15:33 ---
I'll take a look at it.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38228
--- Comment #3 from bonzini at gnu dot org 2009-01-31 15:39 ---
In both versions there's some pessimization in the expansion of _mm_set_ps and
_mm_set_ps1. It's probably easier to fix than the regression.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38682
--- Comment #4 from bonzini at gnu dot org 2009-01-31 16:23 ---
I see optimal code with trunk:
.LFB8:
movaps %xmm1, %xmm4
shrl$2, %edx
mov %edx, %edx
xorl%eax, %eax
addss %xmm0, %xmm4
movaps %xmm4, %xmm3
unpcklps
--- Comment #4 from michal at rzechonek dot net 2009-01-31 16:58 ---
I think that this bug is related to failure in code below:
class movable {
int resource; // 0=not acquired
movable(const movable); // no copy c'tor
void operator=(const movable); // no copy-assignment
--- Comment #24 from ktietz at gcc dot gnu dot org 2009-01-31 17:21 ---
(In reply to comment #21)
Hi Joey, thanks for helping look at this bug.
If you catch up with all the comments, you'll see that it's not just Cygwin,
SjLj was broken on Linux too; the mechanism works the
--- Comment #5 from rguenth at gcc dot gnu dot org 2009-01-31 17:25 ---
Confirmed.
struct S
{
S(): i(2) {}
S(S const);
int i;
};
void f(S x) { x.i = 0; }
extern C void abort (void);
int main()
{
S y;
f(static_castS(y));
if (y.i != 2)
abort ();
return 0;
}
declaring
--- Comment #25 from hjl at gcc dot gnu dot org 2009-01-31 17:34 ---
Subject: Bug 38952
Author: hjl
Date: Sat Jan 31 17:34:03 2009
New Revision: 143823
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=143823
Log:
2009-01-31 Dave Korn dave.korn.cyg...@gmail.com
PR
--- Comment #26 from hjl dot tools at gmail dot com 2009-01-31 17:34
---
Fixed.
--
hjl dot tools at gmail dot com changed:
What|Removed |Added
Status|NEW
--- Comment #9 from hjl dot tools at gmail dot com 2009-01-31 17:43 ---
The code is valid in gnu89 mode.
--
hjl dot tools at gmail dot com changed:
What|Removed |Added
--- Comment #5 from hjl dot tools at gmail dot com 2009-01-31 17:44 ---
Fixed.
--
hjl dot tools at gmail dot com changed:
What|Removed |Added
model: posix
gcc version 4.4.0 20090131 (experimental) [trunk revision 143817] (GCC)
Note: I use GNU's 'ld' on OpenSolaris.
When I run make -i check I get this:
...
Running /usr/share/src/gcc_trunk/gcc/testsuite/gcc.dg/torture/dg-torture.exp
...
FAIL: gcc.dg/torture/fp-int-convert-float128.c -O0
--- Comment #8 from hjl dot tools at gmail dot com 2009-01-31 17:49 ---
Don't we need to add a testcase in comment #3?
--
hjl dot tools at gmail dot com changed:
What|Removed |Added
--- Comment #9 from rguenth at gcc dot gnu dot org 2009-01-31 17:53 ---
Feel free to reduce one. I think using libstdc++ in g++.dg is not possible?
--
rguenth at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #5 from dnovillo at gcc dot gnu dot org 2009-01-31 18:49
---
Subject: Bug 39010
Author: dnovillo
Date: Sat Jan 31 18:49:10 2009
New Revision: 143828
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=143828
Log:
Revert
2009-01-30 H.J. Lu
--- Comment #6 from dnovillo at gcc dot gnu dot org 2009-01-31 18:51
---
The patch
2009-01-30 H.J. Lu hongjiu...@intel.com
PR lto/39010
* lto-function-out.c (output_string): Properly handle the
trailing '\0'.
--- Comment #3 from davek at gcc dot gnu dot org 2009-01-31 18:52 ---
Subject: Bug 38904
Author: davek
Date: Sat Jan 31 18:52:00 2009
New Revision: 143829
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=143829
Log:
PR target/38904
* mkmap-flat.awk (END): Use
--- Comment #4 from dave dot korn dot cygwin at gmail dot com 2009-01-31
18:53 ---
Bug fixed.
--
dave dot korn dot cygwin at gmail dot com changed:
What|Removed |Added
--- Comment #3 from ktietz at gcc dot gnu dot org 2009-01-31 18:56 ---
This case fails for the target x86_64-pc-mingw32 for the same reason. It seems
to be a recursion issue in gimplifier.
On w64 it produces a stack overflow with a call deepth of about #16600 frames.
--
ktietz at
--- Comment #9 from tydeman at tybor dot com 2009-01-31 20:16 ---
4.3.2-7 still has problems (even with -frounding-math).
A more complete test can be found at:
http://www.tybor.com/tflt2int.c
--
tydeman at tybor dot com changed:
What|Removed
20090131 (experimental) [trunk revision 143817] (GCC)
wget -c ftp://ftp.gnu.org/gnu/gforth/gforth-0.7.0.tar.gz
gunzip -d gforth-0.7.0.tar.gz
gtar -xf gforth-0.7.0.tar
export set CC=gcc -ffloat-store -msse3 -mssse3 -m3dnow -msse4.1
-mfpmath=sse,387
export set CXX=g++ -ffloat-store -msse3 -mssse3
--- Comment #3 from H9XLrv5oXVNvHiUI at spambox dot us 2009-01-31 20:28
---
Sorry for the late reply!
I thought I'd receive an e-mail when one of my reports gets updated, silly me.
Yes I know it works if you switch the order, but that's exactly the point of
the code, to have them in
I was giving LTO a play and trying to look at some of the testsuite failures.
One of the failures that I am seeing is an ICE when I try to compile an empty
function. I reduced this from the ICE for c-torture/compile/20021204-1.c
void t(void)
{
}
The ICE occurs in cc1 with and without LTO
--- Comment #1 from hal at oz dot net 2009-01-31 21:00 ---
Still present as of the weekly gcc-4.4-20090130 drop. gcc 4.3.3 builds
successfully on the same system with the same configuration.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38102
--- Comment #8 from rob1weld at aol dot com 2009-01-31 21:01 ---
Sometimes you can skirt around the GENERAL_REGS error with other options.
When I use -O0, -O1, -O2, -O3, -Os I get different results.
Using -O0 produces no screen output _or_ compiled file (thus it 'looks'
as though it
Consider:
struct S {
private:
S(const S); // no copy-c'tor
};
S ref();
void f(S);
void g() {
f(ref());
}
This compiles using
g++ -c -std=c++0x test.cpp
where in fact the compiler should complain about the copy-c'tor not being
public.
--
--- Comment #39 from rguenth at gcc dot gnu dot org 2009-01-31 21:36
---
4.0.4 4.1.2 4.2.4 4.3.3 4.4.0
-O0: 1.641.581.911.862.20
-O1: 4.124.60 10.145.315.43
-O2: 6.106.60 13.348.408.57
-Os: 4.825.38 10.926.54
--- Comment #6 from tkoenig at gcc dot gnu dot org 2009-01-31 22:00 ---
Test case now generates this assembly:
.globl main
.type main, @function
main:
pushl %ebp
movl%esp, %ebp
#APP
# 18 gaga.c 1
#start1
# 0 2
# 20 gaga.c 1
#end1
# 0 2
(rev 143833)
Currently --without-ppl and --without-cloog will disable the use of these
libraries, but the top level configure will still try to do version checking on
them. It would be much cleaner to disable these checks when the libraries are
disabled.
Right now the PPL check returns yes,
--- Comment #45 from rob1weld at aol dot com 2009-01-31 22:13 ---
Here is another attempt using gcc version 4.4.0 20090128 (experimental) [trunk
revision 143729] to compile gcc version 4.4.0 20090131 (experimental) [trunk
revision 143817].
I have an Athlon X2 running OpenSolaris in 32
--- Comment #40 from steven at gcc dot gnu dot org 2009-01-31 22:22 ---
Comparing optimized build times *is* very, very fair as we add optimization
capabilities that help nothing for 99% of the code (cselim, bswap optimization,
cestore, etc.), and with a passes pipeline that
--- Comment #9 from hp at gcc dot gnu dot org 2009-01-31 22:25 ---
(In reply to comment #8)
Can you check comment #6?
Sorry, I'm a bit busy right now, but will as soon as I can.
On the other hand, anyone can check it, so it's not necessary to keep this PR
busy.
--
hp at gcc dot
--- Comment #1 from ubizjak at gmail dot com 2009-01-31 22:28 ---
You can use following (slightly wrapped...) patch to enable 128bit soft-fp on
Solaris:
--cut here--
Index: config.gcc
===
--- config.gcc (revision 143833)
--
hp at gcc dot gnu dot org changed:
What|Removed |Added
Status|NEW |WAITING
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37889
In the test case below, both stores are done.
The compiler chould eliminate the first loop completely.
$ cat a.c
void foo(int *a, int n)
{
int i;
for (i=0; in; i++)
a[i] = 0;
for (i=0; in; i++)
a[i] = 1;
}
$ gcc -O3 -S a.c
$ cat a.s
.file a.c
.text
--- Comment #9 from rob1weld at aol dot com 2009-02-01 00:49 ---
It can be compiled with ONE earlier version of gcc but not another - so this is
a regression.
Works on 4.0.2:
# /opt/csw/gcc4/bin/gcc -v
Reading specs from /opt/csw/gcc4/lib/gcc/i386-pc-solaris2.8/4.0.2/specs
Target:
--- Comment #4 from bangerth at gmail dot com 2009-02-01 05:07 ---
Confirmed indeed, with this (linker) error message:
g/x c++ x.cc
/tmp/ccjPvb3J.o: In function `main':
x.cc:(.text+0x12): undefined reference to `Function()'
collect2: ld returned 1 exit status
This already fails with
--- Comment #1 from bangerth at gmail dot com 2009-02-01 05:18 ---
Confirmed.
--
bangerth at gmail dot com changed:
What|Removed |Added
Status|UNCONFIRMED
--- Comment #10 from hp at gcc dot gnu dot org 2009-02-01 05:54 ---
Ok, I'm on it.
--
hp at gcc dot gnu dot org changed:
What|Removed |Added
52 matches
Mail list logo
