Message boards :
Number crunching :
3nd: Compiling faster Windows client with Intel C++ compiler (and fftw?)
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · 6 · Next
Author | Message |
---|---|
Metod, S56RKO Send message Joined: 27 Sep 02 Posts: 309 Credit: 113,221,277 RAC: 9 |
Ah, well, I'm kinda lazy. If you send me a diff (example or something), I might do it. Metod ... |
ewitte2 Send message Joined: 20 Mar 05 Posts: 17 Credit: 48,588 RAC: 0 |
Is an average of less than 1hr good? ;) http://setiweb.ssl.berkeley.edu/results.php?userid=8078843 Any chance of getting a SSE3 version that works with A64 (San Diego and Venice support SSE3)? |
Ned Slider Send message Joined: 12 Oct 01 Posts: 668 Credit: 4,375,315 RAC: 0 |
Sent you an e-mail Metod with the diff Ben gave to me :) We'll put it in on the next compile we do :) Ned *** My Guide to Compiling Optimised BOINC and SETI Clients *** *** Download Optimised BOINC and SETI Clients for Linux Here *** |
Sir Ulli Send message Joined: 21 Oct 99 Posts: 2246 Credit: 6,136,250 RAC: 0 |
Is an average of less than 1hr good? ;) you mean this Side http://setiweb.ssl.berkeley.edu/results.php?hostid=895190 btw your Results are amazing... :) overclocked this Host, normally times under one Hour are not normal for this Computer... Greetings from Germany NRW Ulli S@h Berkeley's Staff Friends Club m7 © |
Prognatus Send message Joined: 6 Jul 99 Posts: 1600 Credit: 391,546 RAC: 0 |
The Yahoo! site, where these files are located for download, seems to have reached its quota. Could Maverick - or someone else - please make them available through an 2nd site also? Thanks! |
jshenry1963 Send message Joined: 17 Nov 04 Posts: 182 Credit: 68,878 RAC: 0 |
They are also at: http://home.arcor.de/jojo4ubin/seti/ (but would be nice to have a readme file there also so it can be easily known what files are for what) Thanks and keep it up dudes (and dudettes) Thanks, and Keep on crunchin' John Henry KI4JPL Sevierville TN I started with nothing, and I still have some of it left. <img src="http://www.boincstats.com/stats/banner.php?cpid=989478996ebd8eadba8f0809051cdde2"> |
-= Vyper =- Send message Joined: 5 Sep 99 Posts: 1652 Credit: 1,065,191,981 RAC: 2,537 |
Hey I'm worthless of this c++ stuff but is this to anyone out there to start a SETI build to the X86-64 platform? Check here http://www.amd.com/us-en/assets/content_type/DownloadableAssets/dwamd_30887.pdf#search='amd64%20compile%20applications' Regards Vyper _________________________________________________________________________ Addicted to SETI crunching! Founder of GPU Users Group |
Prognatus Send message Joined: 6 Jul 99 Posts: 1600 Credit: 391,546 RAC: 0 |
<blockquote>jshenry1963 said: They are also at: http://home.arcor.de/jojo4ubin/seti/</blockquote> Been there. But they are just for Athlon 64. I'm looking for the others, like for instance P3 or generic. |
gomeyer Send message Joined: 21 May 99 Posts: 488 Credit: 50,370,425 RAC: 0 |
After downloading from http://home.arcor.de/jojo4ubin/seti/ and following directions at the beginning of this thread, (including the backup thank goodness,) BOINC started up by reporting that each WU could not be linked because 4.09 could not be found. [edit it was there, I assume the app_info file told BOINC "you WILL use 4.11" /edit] All WU's were deleted and BOINC wanted to start downloading. Fortunately it couldn't connect immediately and I was able to yank the plug; Restored from backup and all is well again. May I therefore assume that one must begin with an empty queue to do this upgrade? Several message below lead me to believe that BOINC didn't care about the fact that this is ver. 4.11. ???? Thanks and Regards |
Beyond Send message Joined: 11 Jul 99 Posts: 15 Credit: 3,176,621 RAC: 3 |
Will this client work with the new 9nm Sempron parts that support SSE2: setiathome_4.11_windows_jo-athlon64-sse2_1105-1.exe Or does it also HAVE to support 64bit instructions? |
SoNic Send message Joined: 24 Dec 00 Posts: 140 Credit: 2,963,627 RAC: 0 |
Will this client work with the new 9nm Sempron parts that support SSE2: Where did you got the "9nm" Sempron? I know that Intel and AMD is making only 90nm chips :). That client works perfect with a Pentium 4 @ 1.3GHz (with SSE2) - so it is a 32 bit client. |
Blade2005 Send message Joined: 3 Apr 99 Posts: 8 Credit: 26,365,253 RAC: 0 |
Will this client work with the new 9nm Sempron parts that support SSE2: If you are referring to the 90nm Semprons, it does work pretty well. In fact a tad better than the P3-SSE client. |
Digger Send message Joined: 4 Dec 99 Posts: 614 Credit: 21,053 RAC: 0 |
May I therefore assume that one must begin with an empty queue to do this upgrade? Several message below lead me to believe that BOINC didn't care about the fact that this is ver. 4.11. The new 4.11 app will finish up any 4.09 workunits you have remaining, however you must use the correct app_info.xml file to do it. Everyone should really be using this one since it works in both cases, whether you have units to finish up or not. Make sure you put it in the seti@home project directory along with the new 4.11 app and it should work fine. Right-click here, then save it to your disk. Dig |
belldandy from pleiades Send message Joined: 24 Oct 02 Posts: 4 Credit: 81,580 RAC: 0 |
After reading this thread, I also want to try those optimized clients for my CPU which is a Prescott 2.8Ghz with SSE3 support (no 64-bit support). Here is my experience:- First, I make a back-up of my BOINC folder (thanks god I have done that) Then I downloaded app_info.xml from http://www.andyk.de/app_info.xml and put it in /projects/setiweb.ssl.berkeley.edu/ I then downloaded boinc_4.27_windows_jo-Pentium_package_1305-1.rar from http://home.arcor.de/jojo4ubin/seti/ and extracted all the files in it (boinc.exe, boincmgr.exe, boinc.dll, boinc.scr and compiler-switches.txt) into the BOINC base directory. I then runs BOINC for quite some time, but I noticed that my processing time doesn't really decreased from the usual 3-4 hours each WU. Another thing I noticed is the SETI client is still at version 4.09. Is that normal? bell-chan |
thinksnow Send message Joined: 26 Dec 02 Posts: 41 Credit: 505,798 RAC: 0 |
belldandy, I see you used andyk's xml file, so what you are seeing should be teh finishing off of any WU's you already had. My understanding (and experience from this very thing last night) is that they will finish off under 4.09 at the same time as normal, but once you report and refill, you'll bump to 4.11 and times, at least for me, dropped from ~9,500 seconds per to ~5,400 per. Check the difference between 18 May 2005 0:15:52 UTC and 18 May 2005 1:50:14 UTC. |
gomeyer Send message Joined: 21 May 99 Posts: 488 Credit: 50,370,425 RAC: 0 |
May I therefore assume that one must begin with an empty queue to do this upgrade? Several message below lead me to believe that BOINC didn't care about the fact that this is ver. 4.11. Thanks Dig, I'll give it a shot first chance tomorrow. |
thinksnow Send message Joined: 26 Dec 02 Posts: 41 Credit: 505,798 RAC: 0 |
Actually, looking at my results, while my WU times are certainly shorter, my credit per seconds ratio is remaining exactly the same at ~0.003112 credits per second! [b]CPU time...claimed cred...granted cred.......cred/sec.......granted cred/sec[/b] 5,287.97.......15.65.......pending.......0.002959548 5,245.88.......16.33.......pending.......0.003112919 5,202.70.......16.19.......pending.......0.003111846 5,564.89.......17.32.......pending.......0.003112371 5,324.52.......16.57.......pending.......0.003112018 5,382.02.......16.75.......16.75.......0.003112214.......0.003112214 5,595.03.......17.42.......20.14.......0.003113477.......0.003599623 00059.13.......00.18.......00.22.......0.00304414.......0.003720616 5,463.23.......17.01.......pending.......0.003113543 9,266.00.......28.84.......pending.......0.003112454 9,577.58.......29.81.......29.81.......0.003112477.......0.003112477 00058.67.......00.18.......00.18.......0.003068007.......0.003068007 9,597.72.......29.87.......29.87.......0.003112197.......0.003112197 9,676.14.......30.12.......34.28.......0.003112812.......0.003542735 9,681.94.......30.14.......30.14.......0.003113012.......0.003113012 8,300.22.......25.84.......pending.......0.00311317 9,024.14.......28.09.......28.09.......0.003112762.......0.003112762 |
Beyond Send message Joined: 11 Jul 99 Posts: 15 Credit: 3,176,621 RAC: 3 |
> If you are referring to the 90nm Semprons oops, typo... Anyway, I made an app_info.xml that works here for migrating from either: setiathome_4.09_windows_intelx86.exe or setiathome_4.11_windows_intelx86.exe to setiathome_4.11_windows_jo-athlon64-sse2_1105-1.exe You may want to check it for typos... :-) Maybe someone who knows more about this than I do should take a look at it... Anyway, again, it works here... BTW, the 90nm Sempron does work fine with the setiathome_4.11_windows_jo-athlon64-sse2_1105-1.exe client. Download the app_info.zip here: http://www.citilink.com/~beyond/app_info.zip and unzip it... |
thinksnow Send message Joined: 26 Dec 02 Posts: 41 Credit: 505,798 RAC: 0 |
Just to keep things tidy: I understand now why my credit/second figures are remaining the same after reading a bit more. |
rsisto Send message Joined: 30 Jul 03 Posts: 135 Credit: 729,936 RAC: 0 |
After reading this thread, I also want to try those optimized clients for my CPU which is a Prescott 2.8Ghz with SSE3 support (no 64-bit support). The get the reduced processing time you have to get the optimized seti client, which has a name similar to "setiathome_4.11_windows_...". This was not present in the package you downloaded. This is the actual application that does the crunching. Look in this thread or the previous 2 ones for the optimized seti application. This application is the one you should put in the projects/setiweb... directory alongside the xml file. |
©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.