Message boards :
Number crunching :
Lunatics Windows Installer v0.41 Release Notes
Message board moderation
Previous · 1 . . . 13 · 14 · 15 · 16 · 17 · Next
Author | Message |
---|---|
Mike Send message Joined: 17 Feb 01 Posts: 34265 Credit: 79,922,639 RAC: 80 |
Just so the question doesn`t reamain unanswered i repost it here as well. The other thread might get lost. Your NV 260 is a mid range card. So normal settings would be -unroll 10 -ffa_block 6144 -ffa_block_fetch 1536. Given the fact pre Fermi`s can only handle one instance you can increase the settings. Your settings are already improved -unroll 12 -ffa_block 8192 -ffa_block_fetch 4096. You can use -unroll 12 -ffa_block 12288 -ffa_block_fetch 6144. Evenso you could try -unroll 14 to 16 but use at your own risk. Beware if you experience overflow tasks reduce the settings again until you find the sweet spot. As always your milage might vary. I will rework the NV readme until the next release. BTW: You dont need to post in 2 threads. Mike With each crime and every kindness we birth our future. |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
Thanks. Just so the question doesn`t reamain unanswered i repost it here as well. This two messages you mentioned are not the same. BTW, it looks like not all NV card user know that they can set cmdline settings .. * Best regards! :-) * Philip J. Fry, team seti.international founder. * Optimize your PC for higher RAC. * SETI@home needs your help. * |
zoom3+1=4 Send message Joined: 30 Nov 03 Posts: 65801 Credit: 55,293,173 RAC: 49 |
I only wish cuda32 was a bit faster, since only 266.58 WHQL won't downclock My video cards to 405MHz on Windows 7 x64... The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
Which SAHv7 CUDA app should work on a: Quadro FX 570, 256 MiB, regsPerBlock 8192 computeCap 1.1, multiProcs 2 & Quadro NVS 420, 256 MiB, regsPerBlock 8192 computeCap 1.1, multiProcs 1 - because of the very low RAM just the cuda22 app? Maybe the old 185.85 driver could use less RAM? OS: Win8 x64 Thanks. * Best regards! :-) * Philip J. Fry, team seti.international founder. * Optimize your PC for higher RAC. * SETI@home needs your help. * |
arkayn Send message Joined: 14 May 99 Posts: 4438 Credit: 55,006,323 RAC: 0 |
Which SAHv7 CUDA app should work on a: CUDA 3.2 is the best for those machines. |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
I'm in contact with the owner of this machine: http://setiathome.berkeley.edu/show_host_detail.php?hostid=6997950 I thought maybe the cuda22 app could work because IIRC it use low (lowest) card RAM. Or, where is the problem at this PC? * Best regards! :-) * Philip J. Fry, team seti.international founder. * Optimize your PC for higher RAC. * SETI@home needs your help. * |
William Send message Joined: 14 Feb 13 Posts: 2037 Credit: 17,689,662 RAC: 0 |
Lots of overflows - I'd say host problems not app problems. Memory problems show in stderr - there is nothing of that sort showing here. I'd advise to check temps and give the host a good cleanout. Else, GPU-Z is your friend to see how much headroom you have for mem. A person who won't read has no advantage over one who can't read. (Mark Twain) |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
It looks like it wasn't mentioned to now here .. Additional to the apps included in the Installer v0.41: Stock MB v7.00 AKv8c_Bb_r1846_winx86_SSE2x.7z AKv8c_Bb_r1846_winx86_SSSE3x.7z AKv8c_Bb_r1846_winx86_AVXx.7z I saw now there are >additional apps< available: AKv8c_Bb_r1846_winx86_Atom.7z AKv8c_Bb_r1846_winx86_SSE3x.7z AKv8c_Bb_r1846_winx86_SSE41x.7z AKv8c_Bb_r1846_winx86_SSE42x.7z For example: On my Intel Core2 Duo E7600 the SSSE3 was faster than the SSE4.1 (the formerly app for SAHv6). I use now (for SAHv7) the (above mentioned) SSSE3 app, maybe the SSE3 or SSE4.1 would be better/faster? IIRC, in the last Installer was a recommendation which app for which CPU. E.g. SSE3 for Core i3/5/7. This is still true and/or is somewhere a recommendation available for the new apps? BTW, the apps will be included in the next Installer? Thanks. * Best regards! :-) * Philip J. Fry, team seti.international founder. * Optimize your PC for higher RAC. * SETI@home needs your help. * |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
It looks like it wasn't mentioned to now here .. Internal testing at Lunatics was unable to establish any of those additional builds as consistently better than what was put into the installer. But for those willing to take the time to test on theor own system it is quite possible one of those could be better, so I did make them available for download. Future installers could have more or fewer builds included. Joe |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
I have currently the SSSE3 app installed. If I would like to test the SSE4.1 and SSE3 app, they could use the available .wisdom file of the SSSE3 app, or I need to make a 'bench run' (postid=1377434) with two WUs for/with every app? Every app (SSE2 up to AVX) have his own .wisdom file? Thanks. * Best regards! :-) * Philip J. Fry, team seti.international founder. * Optimize your PC for higher RAC. * SETI@home needs your help. * |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
The .wisdom files are the same for all AKv8c rev 1846 builds. Note the file name is formed from the source revision number and the processor name, that's enough. FFTW does its own checking for processor capabilities, so the SIMD level which the app targets does not affect the FFTW wisdom. Joe |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
OK, I edited my app_info.xml file entries and copied the SSE41.exe and SSE41.txt file to the project folder. Started BOINC again and ~ 60 MB WUs failed in a bunch, BOINC Manager wasn't usable (frozen). Then BOINC decided to start an AP WU. So I could intervene, BOINC Manager was again usable. Example, this WU was started with SSSE3 and should continue with the SSE4.1 app: http://setiathome.berkeley.edu/result.php?resultid=3155170605 [EDIT: Exit status -185 (0xffffffffffffff47) ERR_RESULT_START <core_client_version>7.0.64</core_client_version> <![CDATA[ <message> couldn't start app: CreateProcess() failed - (unknown error) </message> ]]>] What I made wrong? The Intel Core2 Duo E7600 can do SSE4.1. The formerly SAHv6 SSE4.1 app worked. Thanks. * Best regards! :-) * Philip J. Fry, team seti.international founder. * Optimize your PC for higher RAC. * SETI@home needs your help. * |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
IIRC, that -185 exit status usually means a missing file, so double check that BOINC didn't delete something. Beyond that I can't guess, maybe you should post your app_info.xml so other eyes can look it over. I did redownload the SSE4.1 package from Lunatics and checked it hasn't been corrupted. And that build ran fine in testing on Claggy's Penryn T8100 system which has the same SIMD capabilities as your Wolfdale E7600. Joe |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
I'm confused now. I D/L again the SSE4.1 .7z file. I used 7-Zip for to unzip. I got a message (in past I didn't read carefully enough, sorry): 'unsupported compression method'. I go to the folder to where I unzipped, all files have 0 byte. I looked to the last unzip folder, also there all files 0 byte. I used the SSE4.1 app which had 0 byte last time. The D/L .7z file have: 893 KB (915.280 Bytes) 896 KB (917.504 Bytes) [size on media] - but 7-Zip can't unzip. Where is the problem? Until now 7-Zip worked fine. Thanks. |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
I unzipped an other .7z file (other/old app), and it worked. So my 7-Zip tool work fine. * Best regards! :-) * Philip J. Fry, team seti.international founder. * Optimize your PC for higher RAC. * SETI@home needs your help. * |
arkayn Send message Joined: 14 May 99 Posts: 4438 Credit: 55,006,323 RAC: 0 |
I'm confused now. Just downloaded and unzipped it fine. What version of 7Zip do you have? |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
I unzipped an other .7z file (other/old app), and it worked. You'll probably need to update to 7-zip 9.04 or later. The change to add LZMA2 was about 4 years ago, IIRC. I'm on dial-up, so try to use such improvements when I become aware of them. Joe |
BilBg Send message Joined: 27 May 07 Posts: 3720 Credit: 9,385,827 RAC: 0 |
I unzipped an other .7z file (other/old app), and it worked. The 'old' .7z files are compressed with LZMA method The 'new' - with LZMA2 So you need 7-Zip 9.20 http://www.7-zip.org/ Â - ALF - "Find out what you don't do well ..... then don't do it!" :) Â |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
Thanks to all. The old installer is for 7-Zip v4.65 and it was used the last time at 2010/05/25. I have now the newest v9.20 installed and the SSE41 .7z file was unzipped - and the PC test now the SSE4.1 app. Thanks. * Best regards! :-) * Philip J. Fry, team seti.international founder. * Optimize your PC for higher RAC. * SETI@home needs your help. * |
BilBg Send message Joined: 27 May 07 Posts: 3720 Credit: 9,385,827 RAC: 0 |
Internal testing at Lunatics was unable to establish any of those additional builds as consistently better than what was put into the installer. But for those willing to take the time to test on their own system it is quite possible one of those could be better, so I did make them available for download. SSE2 vs SSE3 on AMD Athlon II X3 455 For me the SSE3 build is a little slower than SSE2 (the same 30 KB .wisdom file pre-copied to both dirs) May be because the SSE3 implementation in my CPU is not so good, SSE3 have more demand on the RAM bandwidth (my is DDR2-800) or SSE3 build have some non-optimal alignment/order of instructions though I'm sure devs made every effort to align the instructions optimally. Quick timetable WU : PG0009_v7.wu AKv8c_Bb_r1846_winx86_SSE2x.exe -verb -nog : Elapsed 631.500 secs CPU 629.031 secs AKv8c_Bb_r1846_winx86_SSE3x.exe -verb -nog : Elapsed 647.875 secs, speedup: -2.59% ratio: 0.97x CPU 643.500 secs, speedup: -2.30% ratio: 0.98x WU : PG0395_v7.wu AKv8c_Bb_r1846_winx86_SSE2x.exe -verb -nog : Elapsed 630.734 secs CPU 628.672 secs AKv8c_Bb_r1846_winx86_SSE3x.exe -verb -nog : Elapsed 634.953 secs, speedup: -0.67% ratio: 0.99x CPU 632.828 secs, speedup: -0.66% ratio: 0.99x WU : PG0444_v7.wu AKv8c_Bb_r1846_winx86_SSE2x.exe -verb -nog : Elapsed 549.578 secs CPU 546.953 secs AKv8c_Bb_r1846_winx86_SSE3x.exe -verb -nog : Elapsed 552.344 secs, speedup: -0.50% ratio: 0.99x CPU 550.125 secs, speedup: -0.58% ratio: 0.99x WU : PG1327_v7.wu AKv8c_Bb_r1846_winx86_SSE2x.exe -verb -nog : Elapsed 667.063 secs CPU 664.344 secs AKv8c_Bb_r1846_winx86_SSE3x.exe -verb -nog : Elapsed 667.266 secs, speedup: -0.03% ratio: 1.00x CPU 664.563 secs, speedup: -0.03% ratio: 1.00x ------------ CPU: Number of processors 1 Number of cores 3 (max 4) Specification AMD Athlon(tm) II X3 455 Processor Codename Rana Core Speed 3315.6 MHz (16.5 x 200.9 MHz) Core Stepping Technology 45 nm Stock frequency 3300 MHz ------------ Chipset: Northbridge NVIDIA MCP61 rev. A2 Southbridge NVIDIA MCP61 rev. A2 ------------ RAM: Memory Type DDR2 Memory Size 3072 MBytes Memory Frequency 401.9 MHz (1:2) Max bandwidth PC2-6400 (400 MHz) CAS# 5.0 RAS# to CAS# 5 RAS# Precharge 5 Cycle Time (tRAS) 18 ------------ Â - ALF - "Find out what you don't do well ..... then don't do it!" :) Â |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.