|
From: James B. <ja...@ha...> - 2005-12-16 10:28:37
|
Hi, I've run the performance tests on my 4 year old laptop (running an up-to-date Fedora Core 4, gcc 4.0.2), which has a 1.1GHz Pentium III processor - cat /proc/cpuinfo gives model name : Intel(R) Pentium(R) III Mobile CPU 1133MHz After SVN updating to r5348 (vex r1497), the results of make perf are -- Running tests in perf ----------------------------------------- bigcode1 valgrind : 0.5s nl:16.6s (33.1x) mc:23.0s (45.9x) bigcode2 valgrind : 0.5s nl:26.6s (52.2x) mc:42.7s (83.8x) bz2 valgrind : 2.5s nl:14.9s ( 6.0x) mc:55.9s (22.5x) fbench valgrind : 1.8s nl: 5.7s ( 3.2x) mc:26.7s (15.1x) ffbench valgrind : 4.3s nl: 8.6s ( 2.0x) mc:24.6s ( 5.7x) sarp valgrind : 0.1s nl: 1.2s ( 9.1x) mc:30.2s (232.5x) -- Finished tests in perf ----------------------------------------- (snipped slightly to avoid line wrapping) Cheers, James. James Begley -- Telephone +354-575-2000 Marine Research Institute, Skulagata 4, P.O. Box 1390, 121 Reykjavik, Iceland. |
|
From: Nicholas N. <nj...@cs...> - 2005-12-16 16:17:18
|
On Fri, 16 Dec 2005, James Begley wrote: > I've run the performance tests on my 4 year old laptop (running an > up-to-date Fedora Core 4, gcc 4.0.2), which has a 1.1GHz Pentium III > processor - cat /proc/cpuinfo gives > > model name : Intel(R) Pentium(R) III Mobile CPU 1133MHz > > After SVN updating to r5348 (vex r1497), the results of make perf are > -- Running tests in perf ----------------------------------------- > bigcode1 valgrind : 0.5s nl:16.6s (33.1x) mc:23.0s (45.9x) > bigcode2 valgrind : 0.5s nl:26.6s (52.2x) mc:42.7s (83.8x) > bz2 valgrind : 2.5s nl:14.9s ( 6.0x) mc:55.9s (22.5x) > fbench valgrind : 1.8s nl: 5.7s ( 3.2x) mc:26.7s (15.1x) > ffbench valgrind : 4.3s nl: 8.6s ( 2.0x) mc:24.6s ( 5.7x) > sarp valgrind : 0.1s nl: 1.2s ( 9.1x) mc:30.2s (232.5x) > -- Finished tests in perf ----------------------------------------- > (snipped slightly to avoid line wrapping) Thanks for the info. It would be useful to see a comparison between r5348 and 3.1.0 if you have that. You can do it with this command: perl perf/vg_perf --vg=<dir1> --vg=<dir2> perf/ Where <dir1> is the directory holding 3.1.0 and <dir2> is the directory holding r5348. It would also be useful if you could compare the COMPVBITS branch, you can check it out with: svn co svn://www.valgrind.org/valgrind/branches/COMPVBITS and just add a third --vg= option to the command line above. Nick |
|
From: James B. <ja...@ha...> - 2005-12-17 13:14:05
|
Hi, On Fri, 2005-12-16 at 10:17 -0600, Nicholas Nethercote wrote: > Thanks for the info. It would be useful to see a comparison between r5348 > and 3.1.0 if you have that. You can do it with this command: > > perl perf/vg_perf --vg=<dir1> --vg=<dir2> perf/ > > Where <dir1> is the directory holding 3.1.0 and <dir2> is the directory > holding r5348. > > It would also be useful if you could compare the COMPVBITS branch, you can > check it out with: > > svn co svn://www.valgrind.org/valgrind/branches/COMPVBITS > > and just add a third --vg= option to the command line above. OK, I SVN updated to r5363 (vex r1500) re-ran the performance tests to compare V 3.1.0, current SVN and the current COMPVBITS branch. The results are below (again, slightly snipped to avoid line wrapping) -- Running tests in perf ----------------------------------------- -- bigcode1 -- bigcode1 valgrind-3.1.0: 0.5s nl:17.9s (35.8x) mc:25.3s (50.6x) bigcode1 valgrind : 0.5s nl:16.5s (33.0x, 7.8%) mc:23.0s (46.0x, 9.0%) bigcode1 compvbits : 0.5s nl:16.6s (33.2x, 7.5%) mc:22.8s (45.5x, 10.0%) -- bigcode2 -- bigcode2 valgrind-3.1.0: 0.5s nl:28.1s (55.1x) mc:46.1s (90.4x) bigcode2 valgrind : 0.5s nl:26.7s (52.3x, 5.1%) mc:42.8s (83.9x, 7.2%) bigcode2 compvbits : 0.5s nl:26.7s (52.4x, 5.0%) mc:42.8s (83.9x, 7.2%) -- bz2 -- bz2 valgrind-3.1.0: 2.5s nl:16.8s ( 6.8x) mc:58.4s (23.5x) bz2 valgrind : 2.5s nl:15.0s ( 6.0x,11.1%) mc:55.9s (22.5x, 4.3%) bz2 compvbits : 2.5s nl:14.9s ( 6.0x,11.2%) mc:60.9s (24.4x,-4.2%) -- fbench -- fbench valgrind-3.1.0: 1.8s nl: 6.0s ( 3.4x) mc:35.7s (20.2x) fbench valgrind : 1.8s nl: 5.7s ( 3.2x, 5.8%) mc:26.5s (15.0x, 25.8%) fbench compvbits : 1.8s nl: 5.7s ( 3.2x, 5.8%) mc:23.9s (13.5x, 33.1%) -- ffbench -- ffbench valgrind-3.1.0: 4.3s nl: 9.1s ( 2.1x) mc:25.2s ( 5.9x) ffbench valgrind : 4.3s nl: 8.6s ( 2.0x, 4.5%) mc:24.9s ( 5.8x, 1.5%) ffbench compvbits : 4.3s nl: 8.6s ( 2.0x, 4.8%) mc:23.5s ( 5.5x, 7.1%) -- heap -- heap valgrind-3.1.0: 0.6s nl: 8.0s (13.6x) mc:52.6s (89.2x) heap valgrind : 0.6s nl: 6.3s (10.7x,21.1%) mc:52.0s (88.1x, 1.3%) heap compvbits : 0.6s nl: 6.3s (10.6x,21.6%) mc:44.7s (75.7x, 15.1%) -- sarp -- sarp valgrind-3.1.0: 0.1s nl: 1.3s (10.5x) mc:30.3s (252.2x) sarp valgrind : 0.1s nl: 1.2s ( 9.8x, 7.1%) mc:30.1s (251.1x, 0.4%) sarp compvbits : 0.1s nl: 1.2s ( 9.8x, 7.1%) mc:13.8s (114.8x,54.5%) -- Finished tests in perf ----------------------------------------- == 7 programs, 42 timings ================= I hope this helps, Cheers, James. James Begley -- Telephone +354-575-2000 Marine Research Institute, Skulagata 4, P.O. Box 1390, 121 Reykjavik, Iceland. |
|
From: Nicholas N. <nj...@cs...> - 2005-12-19 17:07:05
|
On Sat, 17 Dec 2005, James Begley wrote: > OK, I SVN updated to r5363 (vex r1500) re-ran the performance tests to > compare V 3.1.0, current SVN and the current COMPVBITS branch. The > results are below (again, slightly snipped to avoid line wrapping) > > -- Running tests in perf ----------------------------------------- > -- bigcode1 -- > bigcode1 valgrind-3.1.0: 0.5s nl:17.9s (35.8x) mc:25.3s (50.6x) > bigcode1 valgrind : 0.5s nl:16.5s (33.0x, 7.8%) mc:23.0s (46.0x, 9.0%) > bigcode1 compvbits : 0.5s nl:16.6s (33.2x, 7.5%) mc:22.8s (45.5x, 10.0%) > -- bigcode2 -- > bigcode2 valgrind-3.1.0: 0.5s nl:28.1s (55.1x) mc:46.1s (90.4x) > bigcode2 valgrind : 0.5s nl:26.7s (52.3x, 5.1%) mc:42.8s (83.9x, 7.2%) > bigcode2 compvbits : 0.5s nl:26.7s (52.4x, 5.0%) mc:42.8s (83.9x, 7.2%) > -- bz2 -- > bz2 valgrind-3.1.0: 2.5s nl:16.8s ( 6.8x) mc:58.4s (23.5x) > bz2 valgrind : 2.5s nl:15.0s ( 6.0x,11.1%) mc:55.9s (22.5x, 4.3%) > bz2 compvbits : 2.5s nl:14.9s ( 6.0x,11.2%) mc:60.9s (24.4x,-4.2%) > -- fbench -- > fbench valgrind-3.1.0: 1.8s nl: 6.0s ( 3.4x) mc:35.7s (20.2x) > fbench valgrind : 1.8s nl: 5.7s ( 3.2x, 5.8%) mc:26.5s (15.0x, 25.8%) > fbench compvbits : 1.8s nl: 5.7s ( 3.2x, 5.8%) mc:23.9s (13.5x, 33.1%) > -- ffbench -- > ffbench valgrind-3.1.0: 4.3s nl: 9.1s ( 2.1x) mc:25.2s ( 5.9x) > ffbench valgrind : 4.3s nl: 8.6s ( 2.0x, 4.5%) mc:24.9s ( 5.8x, 1.5%) > ffbench compvbits : 4.3s nl: 8.6s ( 2.0x, 4.8%) mc:23.5s ( 5.5x, 7.1%) > -- heap -- > heap valgrind-3.1.0: 0.6s nl: 8.0s (13.6x) mc:52.6s (89.2x) > heap valgrind : 0.6s nl: 6.3s (10.7x,21.1%) mc:52.0s (88.1x, 1.3%) > heap compvbits : 0.6s nl: 6.3s (10.6x,21.6%) mc:44.7s (75.7x, 15.1%) > -- sarp -- > sarp valgrind-3.1.0: 0.1s nl: 1.3s (10.5x) mc:30.3s (252.2x) > sarp valgrind : 0.1s nl: 1.2s ( 9.8x, 7.1%) mc:30.1s (251.1x, 0.4%) > sarp compvbits : 0.1s nl: 1.2s ( 9.8x, 7.1%) mc:13.8s (114.8x,54.5%) > -- Finished tests in perf ----------------------------------------- Thanks for the data. The bz2 compvbits regression is puzzling, but generally the picture is good. Nick |
|
From: Jeroen N. W. <jn...@xs...> - 2005-12-17 13:20:16
|
The result below was generated on a Pentium III (Katmai). vanilla is rev 5314, VEX rev 1486; trunk is rev 5360, VEX rev 1499; COMPVBITS is rev 5361, VEX rev 1499. 3.1.RC1-inner and COMPVBITS are configured with --enable-inner. Jeroen. -- Running tests in perf ---------------------------------------------- -- bigcode1 -- bigcode1 3.1.RC1-inner: 2.4s nl:52.4s (21.6x, -----) mc:77.2s (31.8x, -----) bigcode1 3.1.RC1-outer: 2.4s nl:51.3s (21.1x, 2.1%) mc:76.7s (31.6x, 0.7%) bigcode1 vanilla: 2.4s nl:50.1s (20.6x, 4.4%) mc:71.9s (29.6x, 7.0%) bigcode1 trunk: 2.4s nl:45.1s (18.5x, 14.0%) mc:67.1s (27.6x, 13.2%) bigcode1 COMPVBITS: 2.4s nl:45.0s (18.5x, 14.2%) mc:66.0s (27.1x, 14.6%) -- bigcode2 -- bigcode2 3.1.RC1-inner: 2.5s nl:92.6s (37.6x, -----) mc:155.4s (63.2x, -----) bigcode2 3.1.RC1-outer: 2.5s nl:91.2s (37.1x, 1.5%) mc:154.7s (62.9x, 0.4%) bigcode2 vanilla: 2.5s nl:83.2s (33.8x, 10.1%) mc:133.9s (54.5x, 13.8%) bigcode2 trunk: 2.5s nl:78.9s (32.1x, 14.8%) mc:130.1s (52.9x, 16.3%) bigcode2 COMPVBITS: 2.5s nl:78.8s (32.0x, 15.0%) mc:129.2s (52.5x, 16.9%) -- bz2 -- bz2 3.1.RC1-inner: 6.5s nl:43.8s ( 6.8x, -----) mc:147.4s (22.8x, -----) bz2 3.1.RC1-outer: 6.5s nl:43.2s ( 6.7x, 1.5%) mc:146.6s (22.6x, 0.6%) bz2 vanilla: 6.5s nl:43.9s ( 6.8x, -0.2%) mc:146.8s (22.7x, 0.4%) bz2 trunk: 6.5s nl:37.9s ( 5.8x, 13.6%) mc:139.5s (21.5x, 5.4%) bz2 COMPVBITS: 6.5s nl:37.7s ( 5.8x, 14.0%) mc:137.2s (21.2x, 6.9%) -- fbench -- fbench 3.1.RC1-inner: 4.2s nl:18.1s ( 4.3x, -----) mc:82.7s (19.6x, -----) fbench 3.1.RC1-outer: 4.2s nl:17.9s ( 4.2x, 1.1%) mc:83.2s (19.7x, -0.6%) fbench vanilla: 4.2s nl:17.4s ( 4.1x, 3.5%) mc:81.4s (19.3x, 1.6%) fbench trunk: 4.2s nl:16.5s ( 3.9x, 8.7%) mc:66.0s (15.6x, 20.2%) fbench COMPVBITS: 4.2s nl:16.4s ( 3.9x, 9.3%) mc:60.8s (14.4x, 26.5%) -- ffbench -- ffbench 3.1.RC1-inner: 6.7s nl:17.4s ( 2.6x, -----) mc:51.2s ( 7.7x, -----) ffbench 3.1.RC1-outer: 6.7s nl:17.4s ( 2.6x, -0.2%) mc:51.4s ( 7.7x, -0.3%) ffbench vanilla: 6.7s nl:17.4s ( 2.6x, 0.0%) mc:51.1s ( 7.7x, 0.1%) ffbench trunk: 6.7s nl:17.0s ( 2.5x, 2.5%) mc:51.0s ( 7.6x, 0.4%) ffbench COMPVBITS: 6.7s nl:17.5s ( 2.6x, -0.3%) mc:47.4s ( 7.1x, 7.5%) -- sarp -- sarp 3.1.RC1-inner: 0.3s nl: 3.3s (10.7x, -----) mc:67.2s (216.7x, -----) sarp 3.1.RC1-outer: 0.3s nl: 3.3s (10.6x, 0.3%) mc:67.0s (216.1x, 0.3%) sarp vanilla: 0.3s nl: 3.2s (10.4x, 3.0%) mc:66.5s (214.4x, 1.1%) sarp trunk: 0.3s nl: 3.0s ( 9.7x, 9.1%) mc:65.9s (212.5x, 1.9%) sarp COMPVBITS: 0.3s nl: 3.0s ( 9.7x, 9.1%) mc:33.9s (109.3x, 49.6%) -- Finished tests in perf ---------------------------------------------- == 6 programs, 60 timings ================= |
|
From: Nicholas N. <nj...@cs...> - 2005-12-19 17:08:26
|
On Sat, 17 Dec 2005, Jeroen N. Witmond wrote: > The result below was generated on a Pentium III (Katmai). vanilla is rev > 5314, VEX rev 1486; trunk is rev 5360, VEX rev 1499; COMPVBITS is rev > 5361, VEX rev 1499. 3.1.RC1-inner and COMPVBITS are configured with > --enable-inner. Generally very encouraging for COMPVBITS, good to see. Thanks. Nick > > Jeroen. > > -- Running tests in perf ---------------------------------------------- > -- bigcode1 -- > bigcode1 3.1.RC1-inner: 2.4s nl:52.4s (21.6x, -----) mc:77.2s (31.8x, -----) > bigcode1 3.1.RC1-outer: 2.4s nl:51.3s (21.1x, 2.1%) mc:76.7s (31.6x, 0.7%) > bigcode1 vanilla: 2.4s nl:50.1s (20.6x, 4.4%) mc:71.9s (29.6x, 7.0%) > bigcode1 trunk: 2.4s nl:45.1s (18.5x, 14.0%) mc:67.1s (27.6x, 13.2%) > bigcode1 COMPVBITS: 2.4s nl:45.0s (18.5x, 14.2%) mc:66.0s (27.1x, 14.6%) > -- bigcode2 -- > bigcode2 3.1.RC1-inner: 2.5s nl:92.6s (37.6x, -----) mc:155.4s (63.2x, > -----) > bigcode2 3.1.RC1-outer: 2.5s nl:91.2s (37.1x, 1.5%) mc:154.7s (62.9x, > 0.4%) > bigcode2 vanilla: 2.5s nl:83.2s (33.8x, 10.1%) mc:133.9s (54.5x, 13.8%) > bigcode2 trunk: 2.5s nl:78.9s (32.1x, 14.8%) mc:130.1s (52.9x, 16.3%) > bigcode2 COMPVBITS: 2.5s nl:78.8s (32.0x, 15.0%) mc:129.2s (52.5x, 16.9%) > -- bz2 -- > bz2 3.1.RC1-inner: 6.5s nl:43.8s ( 6.8x, -----) mc:147.4s (22.8x, > -----) > bz2 3.1.RC1-outer: 6.5s nl:43.2s ( 6.7x, 1.5%) mc:146.6s (22.6x, > 0.6%) > bz2 vanilla: 6.5s nl:43.9s ( 6.8x, -0.2%) mc:146.8s (22.7x, 0.4%) > bz2 trunk: 6.5s nl:37.9s ( 5.8x, 13.6%) mc:139.5s (21.5x, 5.4%) > bz2 COMPVBITS: 6.5s nl:37.7s ( 5.8x, 14.0%) mc:137.2s (21.2x, 6.9%) > -- fbench -- > fbench 3.1.RC1-inner: 4.2s nl:18.1s ( 4.3x, -----) mc:82.7s (19.6x, -----) > fbench 3.1.RC1-outer: 4.2s nl:17.9s ( 4.2x, 1.1%) mc:83.2s (19.7x, -0.6%) > fbench vanilla: 4.2s nl:17.4s ( 4.1x, 3.5%) mc:81.4s (19.3x, 1.6%) > fbench trunk: 4.2s nl:16.5s ( 3.9x, 8.7%) mc:66.0s (15.6x, 20.2%) > fbench COMPVBITS: 4.2s nl:16.4s ( 3.9x, 9.3%) mc:60.8s (14.4x, 26.5%) > -- ffbench -- > ffbench 3.1.RC1-inner: 6.7s nl:17.4s ( 2.6x, -----) mc:51.2s ( 7.7x, -----) > ffbench 3.1.RC1-outer: 6.7s nl:17.4s ( 2.6x, -0.2%) mc:51.4s ( 7.7x, -0.3%) > ffbench vanilla: 6.7s nl:17.4s ( 2.6x, 0.0%) mc:51.1s ( 7.7x, 0.1%) > ffbench trunk: 6.7s nl:17.0s ( 2.5x, 2.5%) mc:51.0s ( 7.6x, 0.4%) > ffbench COMPVBITS: 6.7s nl:17.5s ( 2.6x, -0.3%) mc:47.4s ( 7.1x, 7.5%) > -- sarp -- > sarp 3.1.RC1-inner: 0.3s nl: 3.3s (10.7x, -----) mc:67.2s (216.7x, > -----) > sarp 3.1.RC1-outer: 0.3s nl: 3.3s (10.6x, 0.3%) mc:67.0s (216.1x, > 0.3%) > sarp vanilla: 0.3s nl: 3.2s (10.4x, 3.0%) mc:66.5s (214.4x, 1.1%) > sarp trunk: 0.3s nl: 3.0s ( 9.7x, 9.1%) mc:65.9s (212.5x, 1.9%) > sarp COMPVBITS: 0.3s nl: 3.0s ( 9.7x, 9.1%) mc:33.9s (109.3x, 49.6%) > -- Finished tests in perf ---------------------------------------------- > > == 6 programs, 60 timings ================= > > > > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log files > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click > _______________________________________________ > Valgrind-developers mailing list > Val...@li... > https://lists.sourceforge.net/lists/listinfo/valgrind-developers > |