public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug middle-end/15760] New: Loop unrolling bug with max-unrolled-insns=200
@ 2004-06-01 10:22 micis at gmx dot de
2004-06-01 10:28 ` [Bug middle-end/15760] " micis at gmx dot de
` (3 more replies)
0 siblings, 4 replies; 5+ messages in thread
From: micis at gmx dot de @ 2004-06-01 10:22 UTC (permalink / raw)
To: gcc-bugs
It was stated that in the second prerelease of GCC-3.3.4 from 30 May 2004
the -funroll-loop miscompilation has been fixed.
That is true for the default option set.
However, if you use the options
-O2 -funroll-loops --param max-unrolled-insns=xxx
execution of memcpy-1.c will still fail for all xxx > 192
(I only tested PowerPC target)
On PowerPC (G5) I found that --param max-unrolled-insns=600 results
in about 20% - 30% higher speed in my application (mostly Altivec image
processing).
Michael Cieslinski
--
Summary: Loop unrolling bug with max-unrolled-insns=200
Product: gcc
Version: 3.3.4
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: middle-end
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: micis at gmx dot de
CC: gcc-bugs at gcc dot gnu dot org
GCC build triplet: powerpc-unknown-linux-gnu
GCC host triplet: powerpc-unknown-linux-gnu
GCC target triplet: powerpc-unknown-linux-gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15760
^ permalink raw reply [flat|nested] 5+ messages in thread
* [Bug middle-end/15760] Loop unrolling bug with max-unrolled-insns=200
2004-06-01 10:22 [Bug middle-end/15760] New: Loop unrolling bug with max-unrolled-insns=200 micis at gmx dot de
@ 2004-06-01 10:28 ` micis at gmx dot de
2004-06-01 11:26 ` pinskia at gcc dot gnu dot org
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: micis at gmx dot de @ 2004-06-01 10:28 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From micis at gmx dot de 2004-06-01 10:28 -------
This bug is present in gcc for a long time, at least in gcc 3.3.3
I only issue it here again, because it was stated, that it had been fixed in
the 2nd prerelease of gcc3.3.4
output from gcc -v
Reading specs from /usr/local/gcc33/lib/gcc-lib/powerpc-unknown-linux-
gnu/3.3.4/specs
Configured with: ../gcc33/configure --prefix=/usr/local/gcc33 --program-
suffix=33 --with-cpu=7450 --enable-altivec --enable-languages=c,c++ --with-
fmessage-length=0 --with-falign-loops=16 --with-falign-jumps=16 --with-falign-
functions=32 --with-fstrict-aliasing --with-funroll-loops --with-mpowerpc-
gpopt --with-fsched-interblock --with-fno-gcse-sm :
(reconfigured) ../gcc33/configure --prefix=/usr/local/gcc33 --program-
suffix=33 --with-cpu=7450 --enable-altivec --enable-languages=c,c++
Thread model: posix
gcc version 3.3.4 20040530 (prerelease)
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15760
^ permalink raw reply [flat|nested] 5+ messages in thread
* [Bug middle-end/15760] Loop unrolling bug with max-unrolled-insns=200
2004-06-01 10:22 [Bug middle-end/15760] New: Loop unrolling bug with max-unrolled-insns=200 micis at gmx dot de
2004-06-01 10:28 ` [Bug middle-end/15760] " micis at gmx dot de
@ 2004-06-01 11:26 ` pinskia at gcc dot gnu dot org
2004-06-01 14:20 ` bangerth at dealii dot org
2004-08-14 7:19 ` pinskia at gcc dot gnu dot org
3 siblings, 0 replies; 5+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-06-01 11:26 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From pinskia at gcc dot gnu dot org 2004-06-01 11:26 -------
We need the preprocessed source. read http://gcc.gnu.org/bugs.html.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |WAITING
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15760
^ permalink raw reply [flat|nested] 5+ messages in thread
* [Bug middle-end/15760] Loop unrolling bug with max-unrolled-insns=200
2004-06-01 10:22 [Bug middle-end/15760] New: Loop unrolling bug with max-unrolled-insns=200 micis at gmx dot de
2004-06-01 10:28 ` [Bug middle-end/15760] " micis at gmx dot de
2004-06-01 11:26 ` pinskia at gcc dot gnu dot org
@ 2004-06-01 14:20 ` bangerth at dealii dot org
2004-08-14 7:19 ` pinskia at gcc dot gnu dot org
3 siblings, 0 replies; 5+ messages in thread
From: bangerth at dealii dot org @ 2004-06-01 14:20 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From bangerth at dealii dot org 2004-06-01 14:20 -------
Andrew, this is gcc's own testcase
./gcc/testsuite/gcc.c-torture/execute/memcpy-1.c
Alas, not having a PPC at hand, I can't test this. Can you?
W.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15760
^ permalink raw reply [flat|nested] 5+ messages in thread
* [Bug middle-end/15760] Loop unrolling bug with max-unrolled-insns=200
2004-06-01 10:22 [Bug middle-end/15760] New: Loop unrolling bug with max-unrolled-insns=200 micis at gmx dot de
` (2 preceding siblings ...)
2004-06-01 14:20 ` bangerth at dealii dot org
@ 2004-08-14 7:19 ` pinskia at gcc dot gnu dot org
3 siblings, 0 replies; 5+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-08-14 7:19 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From pinskia at gcc dot gnu dot org 2004-08-14 07:19 -------
I cannot reproduce it at all in 3.3 or 3.4.0 or 3.5.0 so closing as fixed in 3.4.0.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|WAITING |RESOLVED
Resolution| |FIXED
Target Milestone|--- |3.4.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15760
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2004-08-14 7:19 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-06-01 10:22 [Bug middle-end/15760] New: Loop unrolling bug with max-unrolled-insns=200 micis at gmx dot de
2004-06-01 10:28 ` [Bug middle-end/15760] " micis at gmx dot de
2004-06-01 11:26 ` pinskia at gcc dot gnu dot org
2004-06-01 14:20 ` bangerth at dealii dot org
2004-08-14 7:19 ` pinskia 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).