Message boards :
Number crunching :
Contributing code? Amd64 build for Windows
Message board moderation
Author | Message |
---|---|
jstelly Send message Joined: 3 Apr 99 Posts: 5 Credit: 0 RAC: 0 |
I've got boinc and boinc_seti built as AMD64 binaries for Windows XP for Extended Systems. If anyone wants to try out either/both, they seem to be running for me, they run through the sample work unit without error. They're available here: http://js.ofasoft.com/amd64/boinc_seti_amd64.zip That client is just built with standard optimization for amd64. I'm doing some profiling now and going to try and pull out some more of the FFT functions and SSE-ize them. My question is, How do I get the changes I make back to the dev team to be cosnidered for addition to the standard source? Is there a developer area or is there one planned now that the project is over on boinc? |
Sir Ulli Send message Joined: 21 Oct 99 Posts: 2246 Credit: 6,136,250 RAC: 0 |
Thanks a lot for this, a lot of people are waiting for this. one Question, this Version runs only only on MS 64-Bit ore ?, and have you comparisons Times between 32 bit and your 64 bit optimized Version, i will give i a try tomorrow on and Epox 8HDA3+ with an Athlon 64 3.200 at 11 x 215 and memory Twinmos BH5 Chips at 2.0 2 2 7, WD 2000 GIG SATA with 8 MB Cache. Thanka a lot for your hard work. Greetings from Germany NRW Ulli [/url] |
jstelly Send message Joined: 3 Apr 99 Posts: 5 Credit: 0 RAC: 0 |
It only runs on XP for Extended Systems (I built and ran it on build 1218, not sure about earlier builds or Win2k3 for x64). I haven't got around to timing a 32-bit version vs. the 64-bit build. I'd like to get the optimization done and then I figure I'll see where we are. It was built using Visual Studio 2005 beta 1 so I'm not too confident in the optimizing abilities of the compiler, especially for amd64 which is why I want to go through it by hand a bit. If some folks want to try downloading it and running some timing tests, I'd be interested in hearing some results. |
TC Send message Joined: 18 Aug 99 Posts: 4 Credit: 2,673,384 RAC: 0 |
I'll try it at soon as we have something to try it on LOL. Sure hope you can squeeze some more optimizations out of it. I've got two A64's, but they're both running linux at the moment. |
Heffed Send message Joined: 19 Mar 02 Posts: 1856 Credit: 40,736 RAC: 0 |
Any hope of AMD64 optimizations running on a 32 bit platform? (probably too much to hope for) <a> [/url] |
Sir Ulli Send message Joined: 21 Oct 99 Posts: 2246 Credit: 6,136,250 RAC: 0 |
>My question is, How do I get the changes I make back to the dev team to be >cosnidered for addition to the standard source? Is there a developer area or is >there one planned now that the project is over on boinc? I have mailed Athlon Rob for the E-mail of Rom Walton, he is one of the Main programmers at Boinc in Berkeley, if you have questions, my E-mail is ullibrinkschmidt@teleos-web.de thanks for your Work. Greetings from Germany NRW Ulli [/url] |
jstelly Send message Joined: 3 Apr 99 Posts: 5 Credit: 0 RAC: 0 |
> Any hope of AMD64 optimizations running on a 32 bit platform? (probably too > much to hope for) Well, the changes I'm making aren't really amd64 specific, they'd work on any CPU that supports SSE, so I'll build a 32-bit binary with the changes and see if they help performance any. I don't know what the current setup is for building the "official" seti client binaries. If they're being built with the Intel Compiler then these changes might not help much there (the Intel compiler is pretty good at optimizing 32-bit code). |
JigPu Send message Joined: 16 Feb 00 Posts: 99 Credit: 2,513,738 RAC: 0 |
Good to hear! I've been wanting to see if anybody would try to optimize the Windows builds now that the source is released, and I wish you luck! I would try optimizing myself, except for a lack of descent compiler and the fact that I'm still n00bish enough at C++ that I'd get lost wading through the source :D _______________________________ |
Sir Ulli Send message Joined: 21 Oct 99 Posts: 2246 Credit: 6,136,250 RAC: 0 |
hmmm, i am no be able to get this running, but starting the CLI i get this Benchmarks with the 64-Bit Client --- - 2004-08-22 15:43:56 - Benchmark results: --- - 2004-08-22 15:43:56 - Number of CPUs: 1 --- - 2004-08-22 15:43:56 - 3118 double precision MIPS (Whetstone) per CPU --- - 2004-08-22 15:43:56 - 25645 integer MIPS (Dhrystone) per CPU with the normal Windows Client --- - 2004-08-22 15:30:56 - Benchmark results: --- - 2004-08-22 15:30:56 - Number of CPUs: 1 --- - 2004-08-22 15:30:56 - 2452 double precision MIPS (Whetstone) per CPU --- - 2004-08-22 15:30:56 - 5045 integer MIPS (Dhrystone) per CPU That is amasing @jstelly any Tipps who i can get this running Greetings from Germany NRW Ulli [/url] |
jstelly Send message Joined: 3 Apr 99 Posts: 5 Credit: 0 RAC: 0 |
What exactly is happening or isn't happening? The install for this would be: 1) Extract the contents of the zip file to a directory (this will be your boinc directory) 2) Run boinc_cli.exe 3) Enter "http://setiathome.berkeley.edu" to attach to the SETI project. 4) Enter your SETI id key that you got in the email after registering with the project. 5) The boinc client should download work units (when the servers are functioning) and kick off the seti client. Which of those steps aren't working? |
Sir Ulli Send message Joined: 21 Oct 99 Posts: 2246 Credit: 6,136,250 RAC: 0 |
the Servers at Borkeley are down since 13:00 UTC Time, i will try this tomorrow again. Greetings from Germany NRW Ulli [/url] |
slavko.sk Send message Joined: 27 Jun 00 Posts: 346 Credit: 417,028 RAC: 0 |
I'm very interested in 64 bits client. I go to buy Athlon64 this week and install Win64 for ES on it. Try WAP user info http://slavko.sk/wap/seti |
STE\/E Send message Joined: 29 Mar 03 Posts: 1137 Credit: 5,334,063 RAC: 0 |
3) Enter "http://setiathome.berkeley.edu" to attach to the SETI project. ========== I don't believe that is the correct URL Address to be ataching to..The Address I use and have no problem with is "http://setiweb.ssl.berekely.edu/" Try that one & see what happens ... |
slavko.sk Send message Joined: 27 Jun 00 Posts: 346 Credit: 417,028 RAC: 0 |
http://setiathome.berkeley.edu is right one but yes, it works with http://setiweb.ssl.berekely.edu/ too. > 3) Enter "http://setiathome.berkeley.edu" to attach to the SETI project. > ========== > > I don't believe that is the correct URL Address to be ataching to..The Address > I use and have no problem with is "http://setiweb.ssl.berekely.edu/" > > Try that one & see what happens ... Try WAP user info http://slavko.sk/wap/seti |
Keck_Komputers Send message Joined: 4 Jul 99 Posts: 1575 Credit: 4,152,111 RAC: 1 |
> 3) Enter "http://setiathome.berkeley.edu" to attach to the SETI project. > ========== > > I don't believe that is the correct URL Address to be ataching to..The Address > I use and have no problem with is "http[snip]/" > > Try that one & see what happens ... > > http://setiathome.berkeley.edu is the official master URL for the seti project. There are a couple of other addresses that may work temporarily. If you are attached to any URL other than http://setiathome.berkeley.edu you will eventually have to detach and reattach when the address you are using no longer has the needed information to run the project. John Keck BOINCing since 2002/12/08 |
Sir Ulli Send message Joined: 21 Oct 99 Posts: 2246 Credit: 6,136,250 RAC: 0 |
> What exactly is happening or isn't happening? The install for this would be: > > 1) Extract the contents of the zip file to a directory (this will be your > boinc directory) > > 2) Run boinc_cli.exe > > 3) Enter "http://setiathome.berkeley.edu" to attach to the SETI project. > > 4) Enter your SETI id key that you got in the email after registering with the > project. > > 5) The boinc client should download work units (when the servers are > functioning) and kick off the seti client. > > > Which of those steps aren't working? > > all steps are working, but the client cant connect to Berkeley 2004-08-25 18:45:11 [---] Can't resolve hostname setiathome.berkeley.edu 2004-08-25 18:45:11 [---] Can't resolve hostname setiathome.berkeley.edu 2004-08-25 18:45:11 [SETI@home] Couldn't read master page for SETI@home: error - 113 2004-08-25 18:45:11 [SETI@home] Couldn't read master page for SETI@home: error - 113 2004-08-25 18:45:11 [SETI@home] Master file fetch failed 2004-08-25 18:45:11 [SETI@home] Master file fetch failed i have a lan connection with no Proxy, so the normal 32 Bit client has no Probs. Greetings from Germany NRW Ulli [/url] |
Frank Uittenbosch Send message Joined: 6 Jul 00 Posts: 10 Credit: 5,667,886 RAC: 0 |
Hello guys, the download location no longer seems to be working, is there an alternative ? and will it work and all with the new version of boinc ? Hobbies: Http://www.memleak.nl - H2o cooling - case modding - PHP developing - space - science - IT - and every possible way to help mankind find Alien Lifeforms |
Sir Ulli Send message Joined: 21 Oct 99 Posts: 2246 Credit: 6,136,250 RAC: 0 |
that Version was never working, ... Benchmarks are super, but that was all.. I here the guys at TPR are looking for an working 64-Bit Version, and the Guys at L'Alliance Francophone are doing the same, but to this Time, i think it is still in Progress. Greetings from Germany NRW Ulli [/url] |
Frank Uittenbosch Send message Joined: 6 Jul 00 Posts: 10 Credit: 5,667,886 RAC: 0 |
ok, thx, let's yust hope it'll be there soon :) so we can crunch more :) Hobbies: Http://www.memleak.nl - H2o cooling - case modding - PHP developing - space - science - IT - and every possible way to help mankind find Alien Lifeforms |
slavko.sk Send message Joined: 27 Jun 00 Posts: 346 Credit: 417,028 RAC: 0 |
So when I have 2724 / 5431 benchmarks on my Athlon 64 3200+ @ 32bit client I may expect 3464 / 27607 @ 64 bit client??? It is real movement? It is not a joke? Then it would be great. And I'm running in defaults till now. But I have bought AS3 today (nobody sells AS5 in Slovakia) and ThermalTake case Fun to start tweaking my PC and get more power. (I'm watching this thread and waiting for 64 bits client for Win64) > hmmm, i am no be able to get this running, but starting the CLI i get this > Benchmarks with the 64-Bit Client > > --- - 2004-08-22 15:43:56 - Benchmark results: > --- - 2004-08-22 15:43:56 - Number of CPUs: 1 > --- - 2004-08-22 15:43:56 - 3118 double precision MIPS (Whetstone) per CPU > --- - 2004-08-22 15:43:56 - 25645 integer MIPS (Dhrystone) per CPU > > with the normal Windows Client > > --- - 2004-08-22 15:30:56 - Benchmark results: > --- - 2004-08-22 15:30:56 - Number of CPUs: 1 > --- - 2004-08-22 15:30:56 - 2452 double precision MIPS (Whetstone) per CPU > --- - 2004-08-22 15:30:56 - 5045 integer MIPS (Dhrystone) per CPU S@h Berkeley's Staff Friends Club © member |
©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.