public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug rtl-optimization/37397] New: IRA performance impact on SPEC CPU 2K/2006
@ 2008-09-06 15:36 hjl dot tools at gmail dot com
2008-09-06 15:50 ` [Bug rtl-optimization/37397] " hjl dot tools at gmail dot com
` (6 more replies)
0 siblings, 7 replies; 8+ messages in thread
From: hjl dot tools at gmail dot com @ 2008-09-06 15:36 UTC (permalink / raw)
To: gcc-bugs
On Core 2, I am using gcc 4.4 revision 139589 as baseline.
On 64bit, I compare
O2: -O2 -ffast-math
O3: -O3 -ffast-math -funroll-loops
On 32bit, I compare
O2: -O2 -msse2 -mfpmath=sse -ffast-math
O3: -O3 -msse2 -mfpmath=sse -ffast-math -funroll-loops
--
Summary: IRA performance impact on SPEC CPU 2K/2006
Product: gcc
Version: 4.4.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: rtl-optimization
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: hjl dot tools at gmail dot com
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37397
^ permalink raw reply [flat|nested] 8+ messages in thread* [Bug rtl-optimization/37397] IRA performance impact on SPEC CPU 2K/2006
2008-09-06 15:36 [Bug rtl-optimization/37397] New: IRA performance impact on SPEC CPU 2K/2006 hjl dot tools at gmail dot com
@ 2008-09-06 15:50 ` hjl dot tools at gmail dot com
2008-11-11 0:05 ` hjl at gcc dot gnu dot org
` (5 subsequent siblings)
6 siblings, 0 replies; 8+ messages in thread
From: hjl dot tools at gmail dot com @ 2008-09-06 15:50 UTC (permalink / raw)
To: gcc-bugs
------- Comment #1 from hjl dot tools at gmail dot com 2008-09-06 15:49 -------
ira-merge banch at revision 140034 against trunk at revision 139589:
32bit O3 IRA against baseline
164.gzip 0.344353%
175.vpr -2.90615%
176.gcc 4.85658%
181.mcf -0.0595948%
186.crafty 0.277906%
197.parser 1.21396%
252.eon 4.9764%
253.perlbmk 4.35907%
254.gap -6.78161%
255.vortex -1.82224%
256.bzip2 -1.99685%
300.twolf 3.05206%
SPECint_base2000 0.399723%
168.wupwise 8.07107%
171.swim 3.2207%
172.mgrid -4.46097%
173.applu -0.689127%
177.mesa 0.938628%
178.galgel 3.69037%
179.art -0.526015%
183.equake 4.69726%
187.facerec -1.32431%
188.ammp 0.784621%
189.lucas 5.99566%
191.fma3d 0.257448%
200.sixtrack 0.847458%
301.apsi 3.85583%
SPECfp_base2000 1.76232%
400.perlbench 12.3077%
401.bzip2 -8.57143%
403.gcc 3.73444%
429.mcf 0.970874%
445.gobmk 2.84091%
456.hmmer 0%
458.sjeng -1.02041%
462.libquantum -12.7596%
464.h264ref 8.03213%
471.omnetpp -1.0929%
473.astar 4.65116%
483.xalancbmk -0.840336%
SPECint(R)_base2006 0.490196%
410.bwaves -2.64317%
416.gamess 0.621118%
433.milc 2.03046%
434.zeusmp -1.1236%
435.gromacs -3.73832%
436.cactusADM 5.12821%
437.leslie3d 3.7037%
444.namd -7.48299%
447.dealII 4.8583%
450.soplex 2.18182%
453.povray 4.81283%
454.calculix 6.97674%
459.GemsFDTD 4.31655%
465.tonto 4.54545%
470.lbm 13.4078%
481.wrf 2%
482.sphinx3 0%
SPECfp(R)_base2006 2.28571%
64bit O3 IRA against baseline
164.gzip 1.02286%
175.vpr 0.566305%
176.gcc 5.66102%
181.mcf 1.95165%
186.crafty 1.42327%
197.parser -0.762112%
252.eon 0.396221%
253.perlbmk 2.11715%
254.gap 3.33674%
255.vortex 0.685136%
256.bzip2 -0.52506%
300.twolf -0.0715052%
SPECint_base2000 1.30387%
168.wupwise 1.21534%
171.swim -0.703341%
172.mgrid -1.74891%
173.applu -0.337838%
177.mesa -1.44665%
178.galgel -0.690985%
179.art 0.534315%
183.equake 0.417343%
187.facerec -2.43478%
188.ammp 1.57632%
189.lucas 4.48487%
191.fma3d 4.04355%
200.sixtrack 0.265722%
301.apsi 2.46505%
SPECfp_base2000 0.527712%
400.perlbench -2.72727%
401.bzip2 1.1976%
403.gcc 0%
429.mcf 0.431034%
445.gobmk -3.53535%
456.hmmer -5.88235%
458.sjeng -0.925926%
462.libquantum 0%
464.h264ref 4.9505%
471.omnetpp -0.561798%
473.astar 0%
483.xalancbmk -0.760456%
SPECint(R)_base2006 -0.881057%
410.bwaves 0.735294%
416.gamess 2.07254%
433.milc -0.473934%
434.zeusmp -3.06122%
435.gromacs -3.27869%
436.cactusADM 2.25564%
437.leslie3d -2.31214%
444.namd 0%
447.dealII -3.77358%
450.soplex 0%
453.povray -1.76991%
454.calculix -5.42169%
459.GemsFDTD 0%
465.tonto 0.549451%
470.lbm 1.44231%
481.wrf -0.617284%
482.sphinx3 -1.67224%
SPECfp(R)_base2006 -1.00503%
32bit O2 IRA against baseline
164.gzip 0.345304%
175.vpr -0.0938527%
176.gcc 0.27933%
181.mcf 2.91467%
186.crafty 0%
197.parser -0.0626174%
252.eon 0%
253.perlbmk -0.748045%
254.gap -0.0433276%
255.vortex -0.0413736%
256.bzip2 0.29703%
300.twolf 0%
SPECint_base2000 0.234388%
168.wupwise -0.728266%
171.swim -0.355297%
172.mgrid -0.176056%
173.applu -1.36418%
177.mesa 0.995671%
178.galgel -0.252294%
179.art 0%
183.equake 0.0363504%
187.facerec -0.93622%
188.ammp 0.0566893%
189.lucas 0.133111%
191.fma3d 0.846614%
200.sixtrack 0%
301.apsi 0.0937647%
SPECfp_base2000 -0.120108%
400.perlbench -0.487805%
401.bzip2 0%
403.gcc 0%
429.mcf -0.377358%
445.gobmk 0%
456.hmmer 0%
458.sjeng 0.581395%
462.libquantum 0%
464.h264ref 0.438596%
471.omnetpp 0%
473.astar 0%
483.xalancbmk 0%
SPECint(R)_base2006 0%
410.bwaves 0%
416.gamess 0%
433.milc -0.819672%
434.zeusmp 0%
435.gromacs 0%
436.cactusADM -0.970874%
437.leslie3d 0%
444.namd 0%
447.dealII 0%
450.soplex -0.529101%
453.povray -1.27389%
454.calculix 0.147493%
459.GemsFDTD 0%
465.tonto 0%
470.lbm 0.543478%
481.wrf -0.909091%
482.sphinx3 0%
SPECfp(R)_base2006 -0.714286%
64bit O2 IRA against baseline
164.gzip 0.125235%
175.vpr -0.0498008%
176.gcc -0.0726216%
181.mcf -0.428725%
186.crafty 0%
197.parser -0.0617284%
252.eon -0.0644745%
253.perlbmk -1.11562%
254.gap 0.0798403%
255.vortex -0.0327976%
256.bzip2 -0.0856164%
300.twolf 0%
SPECint_base2000 -0.139159%
168.wupwise 0.0404858%
171.swim -0.161394%
172.mgrid 0.308325%
173.applu 0.516944%
177.mesa -0.0342466%
178.galgel -0.328084%
179.art 0.0218771%
183.equake 0.35727%
187.facerec 0.209314%
188.ammp 0.112191%
189.lucas -0.0878735%
191.fma3d 0.165563%
200.sixtrack 0%
301.apsi 0.171821%
SPECfp_base2000 0.0932134%
400.perlbench 0%
401.bzip2 0.568182%
403.gcc 0%
429.mcf 0.515464%
445.gobmk 0%
456.hmmer 0%
458.sjeng 0%
462.libquantum 0%
464.h264ref 0%
471.omnetpp -0.70922%
473.astar 0%
483.xalancbmk 0%
SPECint(R)_base2006 0%
410.bwaves 0%
416.gamess -0.578035%
433.milc 0%
434.zeusmp 0%
435.gromacs 0%
436.cactusADM 0%
437.leslie3d 0.70922%
444.namd 0%
447.dealII 0.3861%
450.soplex 10.241%
453.povray 0.465116%
454.calculix 0%
459.GemsFDTD 0%
465.tonto 0%
470.lbm 3.20856%
481.wrf 0%
482.sphinx3 0%
SPECfp(R)_base2006 0.649351%
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37397
^ permalink raw reply [flat|nested] 8+ messages in thread* [Bug rtl-optimization/37397] IRA performance impact on SPEC CPU 2K/2006
2008-09-06 15:36 [Bug rtl-optimization/37397] New: IRA performance impact on SPEC CPU 2K/2006 hjl dot tools at gmail dot com
2008-09-06 15:50 ` [Bug rtl-optimization/37397] " hjl dot tools at gmail dot com
@ 2008-11-11 0:05 ` hjl at gcc dot gnu dot org
2008-11-14 16:45 ` vmakarov at gcc dot gnu dot org
` (4 subsequent siblings)
6 siblings, 0 replies; 8+ messages in thread
From: hjl at gcc dot gnu dot org @ 2008-11-11 0:05 UTC (permalink / raw)
To: gcc-bugs
------- Comment #2 from hjl at gcc dot gnu dot org 2008-11-11 00:03 -------
Subject: Bug 37397
Author: hjl
Date: Tue Nov 11 00:02:20 2008
New Revision: 141757
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=141757
Log:
2008-11-10 Vladimir Makarov <vmakarov@redhat.com>
PR rtl-optimization/37397
* ira-int.h (struct ira_allocno): New member bad_spill_p.
(ALLOCNO_BAD_SPILL_P): New macro.
* ira-color.c (push_allocnos_to_stack): Check ALLOCNO_BAD_SPILL_P.
* ira-build.c (ira_create_allocno): Initialize
ALLOCNO_BAD_SPILL_P.
(create_cap_allocno, propagate_allocno_info,
remove_unnecessary_allocnos): Set up or update
ALLOCNO_BAD_SPILL_P.
(update_bad_spill_attribute): New function.
(ira_build): Call it.
* ira-costs.c (record_reg_classes): Set up ALLOCNO_BAD_SPILL_P.
Modified:
branches/ira-merge/gcc/ChangeLog.ira
branches/ira-merge/gcc/ira-build.c
branches/ira-merge/gcc/ira-color.c
branches/ira-merge/gcc/ira-costs.c
branches/ira-merge/gcc/ira-int.h
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37397
^ permalink raw reply [flat|nested] 8+ messages in thread* [Bug rtl-optimization/37397] IRA performance impact on SPEC CPU 2K/2006
2008-09-06 15:36 [Bug rtl-optimization/37397] New: IRA performance impact on SPEC CPU 2K/2006 hjl dot tools at gmail dot com
2008-09-06 15:50 ` [Bug rtl-optimization/37397] " hjl dot tools at gmail dot com
2008-11-11 0:05 ` hjl at gcc dot gnu dot org
@ 2008-11-14 16:45 ` vmakarov at gcc dot gnu dot org
2008-11-17 14:32 ` hjl dot tools at gmail dot com
` (3 subsequent siblings)
6 siblings, 0 replies; 8+ messages in thread
From: vmakarov at gcc dot gnu dot org @ 2008-11-14 16:45 UTC (permalink / raw)
To: gcc-bugs
------- Comment #3 from vmakarov at gcc dot gnu dot org 2008-11-14 16:43 -------
Subject: Bug 37397
Author: vmakarov
Date: Fri Nov 14 16:41:56 2008
New Revision: 141860
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=141860
Log:
2008-11-14 Vladimir Makarov <vmakarov@redhat.com>
PR rtl-optimization/37397
* ira-int.h (struct ira_allocno): New member bad_spill_p.
(ALLOCNO_BAD_SPILL_P): New macro.
* ira-color.c (push_allocnos_to_stack): Check ALLOCNO_BAD_SPILL_P.
* ira-build.c (ira_create_allocno): Initialize
ALLOCNO_BAD_SPILL_P.
(create_cap_allocno, propagate_allocno_info,
remove_unnecessary_allocnos): Set up or update
ALLOCNO_BAD_SPILL_P.
(update_bad_spill_attribute): New function.
(ira_build): Call it.
* ira-costs.c (record_reg_classes): Set up ALLOCNO_BAD_SPILL_P.
Modified:
trunk/gcc/ChangeLog
trunk/gcc/ira-build.c
trunk/gcc/ira-color.c
trunk/gcc/ira-costs.c
trunk/gcc/ira-int.h
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37397
^ permalink raw reply [flat|nested] 8+ messages in thread* [Bug rtl-optimization/37397] IRA performance impact on SPEC CPU 2K/2006
2008-09-06 15:36 [Bug rtl-optimization/37397] New: IRA performance impact on SPEC CPU 2K/2006 hjl dot tools at gmail dot com
` (2 preceding siblings ...)
2008-11-14 16:45 ` vmakarov at gcc dot gnu dot org
@ 2008-11-17 14:32 ` hjl dot tools at gmail dot com
2008-12-25 2:38 ` [Bug rtl-optimization/37397] [4.4 Regression] " pinskia at gcc dot gnu dot org
` (2 subsequent siblings)
6 siblings, 0 replies; 8+ messages in thread
From: hjl dot tools at gmail dot com @ 2008-11-17 14:32 UTC (permalink / raw)
To: gcc-bugs
------- Comment #4 from hjl dot tools at gmail dot com 2008-11-17 14:31 -------
Revision 141860 caused 30% slowdown on 454.calculix in SPEC CPU 2006
with -O2 -ffast-math on Linux/Intel64.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37397
^ permalink raw reply [flat|nested] 8+ messages in thread* [Bug rtl-optimization/37397] [4.4 Regression] IRA performance impact on SPEC CPU 2K/2006
2008-09-06 15:36 [Bug rtl-optimization/37397] New: IRA performance impact on SPEC CPU 2K/2006 hjl dot tools at gmail dot com
` (3 preceding siblings ...)
2008-11-17 14:32 ` hjl dot tools at gmail dot com
@ 2008-12-25 2:38 ` pinskia at gcc dot gnu dot org
2008-12-30 2:52 ` Joey dot ye at intel dot com
2008-12-30 3:05 ` pinskia at gcc dot gnu dot org
6 siblings, 0 replies; 8+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2008-12-25 2:38 UTC (permalink / raw)
To: gcc-bugs
------- Comment #5 from pinskia at gcc dot gnu dot org 2008-12-25 02:36 -------
Has this been fixed now?
--
pinskia at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |missed-optimization, ra
Summary|IRA performance impact on |[4.4 Regression] IRA
|SPEC CPU 2K/2006 |performance impact on SPEC
| |CPU 2K/2006
Target Milestone|--- |4.4.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37397
^ permalink raw reply [flat|nested] 8+ messages in thread* [Bug rtl-optimization/37397] [4.4 Regression] IRA performance impact on SPEC CPU 2K/2006
2008-09-06 15:36 [Bug rtl-optimization/37397] New: IRA performance impact on SPEC CPU 2K/2006 hjl dot tools at gmail dot com
` (4 preceding siblings ...)
2008-12-25 2:38 ` [Bug rtl-optimization/37397] [4.4 Regression] " pinskia at gcc dot gnu dot org
@ 2008-12-30 2:52 ` Joey dot ye at intel dot com
2008-12-30 3:05 ` pinskia at gcc dot gnu dot org
6 siblings, 0 replies; 8+ messages in thread
From: Joey dot ye at intel dot com @ 2008-12-30 2:52 UTC (permalink / raw)
To: gcc-bugs
------- Comment #6 from Joey dot ye at intel dot com 2008-12-30 02:50 -------
(In reply to comment #4)
> Revision 141860 caused 30% slowdown on 454.calculix in SPEC CPU 2006
> with -O2 -ffast-math on Linux/Intel64.
This regression has been fixed in some revision between 142187 and 142212.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37397
^ permalink raw reply [flat|nested] 8+ messages in thread* [Bug rtl-optimization/37397] [4.4 Regression] IRA performance impact on SPEC CPU 2K/2006
2008-09-06 15:36 [Bug rtl-optimization/37397] New: IRA performance impact on SPEC CPU 2K/2006 hjl dot tools at gmail dot com
` (5 preceding siblings ...)
2008-12-30 2:52 ` Joey dot ye at intel dot com
@ 2008-12-30 3:05 ` pinskia at gcc dot gnu dot org
6 siblings, 0 replies; 8+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2008-12-30 3:05 UTC (permalink / raw)
To: gcc-bugs
------- Comment #7 from pinskia at gcc dot gnu dot org 2008-12-30 03:03 -------
Fixed.
--
pinskia at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |RESOLVED
Resolution| |FIXED
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37397
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2008-12-30 3:05 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2008-09-06 15:36 [Bug rtl-optimization/37397] New: IRA performance impact on SPEC CPU 2K/2006 hjl dot tools at gmail dot com
2008-09-06 15:50 ` [Bug rtl-optimization/37397] " hjl dot tools at gmail dot com
2008-11-11 0:05 ` hjl at gcc dot gnu dot org
2008-11-14 16:45 ` vmakarov at gcc dot gnu dot org
2008-11-17 14:32 ` hjl dot tools at gmail dot com
2008-12-25 2:38 ` [Bug rtl-optimization/37397] [4.4 Regression] " pinskia at gcc dot gnu dot org
2008-12-30 2:52 ` Joey dot ye at intel dot com
2008-12-30 3:05 ` 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).