Posts by Anthony Waters

1) Message boards : Number crunching : No Milkyway with GTX480 (Message 39017) Posted 22 Apr 2010 by Anthony Waters Post: I posted new applications on Milkyway@Home version 3, I only have a GTX 285 so I am not able to test whether it works or not.
2) Message boards : Number crunching : No Milkyway with GTX480 (Message 38747) Posted 15 Apr 2010 by Anthony Waters Post: Correct, the code specifically asks for compute capability 1.3. I will have to look into the changes that compute capability 2.0 brought and determine whether or not they support the CUDA Framework 2.3.
3) Message boards : News : milkyway3 v0.02 source released (Message 38566) Posted 11 Apr 2010 by Anthony Waters Post: Large 11 CUDA new: -3.125086817279229 stock app: -3.125086817277920 ATI3 0.24: -3.125086817279327 12 CUDA new: -3.220820423937230 stock app: -3.220820423936994 ATI3 0.24: -3.220820423937265 20 CUDA new: -2.985310834913899 stock app: -2.985310834912030 ATI3 0.24: -2.985310834913758 21 CUDA new: -2.889340482227060 stock app: -2.889340482223516 ATI3 0.24: -2.889340482226973 79 CUDA new: -2.946681292507908 stock app: -2.946681292501331 ATI3 0.24: -2.946681292507772 82 CUDA new: -2.985567797360019 stock app: -2.985567797346584 ATI3 0.24: -2.985567797359872 86 CUDA new: -3.027907173993949 stock app: -3.027907173971970 ATI3 0.24: -3.027907173993795 CUDA version has the added kahan summation
4) Message boards : Number crunching : Amount of concurrent blocks (Message 38264) Posted 7 Apr 2010 by Anthony Waters Post: The field concurrent blocks dictates how much of the GPU will be used by the application in a single iteration. The more concurrent blocks the faster the workunits are finished, but it will cause more user interface lag.
5) Message boards : Application Code Discussion : CUDA 0.25 Released (Message 36244) Posted 3 Feb 2010 by Anthony Waters Post: Edit: It appears to be returning invalid results therefore it has been temporarily deprecated. Any word on this? A lot of the code was modified to make it faster so I'm tracking down the issue currently. Edit: The issue was identified and the application is going through QA at the moment.
6) Message boards : Application Code Discussion : Recompiled Linux 32/64 apps (Message 36184) Posted 31 Jan 2010 by Anthony Waters Post: PLEASE, PLEASE, PLEASE!!! make ATI GPU 0.20 app for linux x64. I'm pissed off using windows... Right now I'm gonna build a rig for MW, but if there is a chance to avoid this - it will perfect :-) And furthermore - there is app for linux in Collatz. What's the problem to make the same trick for MW? Please see http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=1475, once the code is available I plan on attempting to compile a version for GNU/Linux.
7) Message boards : Application Code Discussion : CUDA 0.25 Released (Message 36034) Posted 25 Jan 2010 by Anthony Waters Post: CUDA 0.25 has been released for 64 bit GNU/Linux, the Windows version will follow once all of the new functionality is verified to be working correctly. Changes: Implemented the ability to read the project specific preference for the number of allowed concurrent blocks. The default value is 128 and allows for very little interface lag at the expense of longer running times. A value of 20000 and above will cut the execution time by as much as 2 minutes, but resulting in noticeable user interface lag. Performance improvements through the use of manual loop unrolling, generating about a 5% increase in performance. Made the final likelihood calculation more parallel, resulting in a decrease of about 3 seconds in execution time. General code cleanup. http://milkyway.cs.rpi.edu/milkyway/download/code_release/mw_gpu_v0.25.tar.gz http://milkyway.cs.rpi.edu/milkyway/download/code_release/mw_gpu_v0.25.zip The latest work units, ps_s222_pes_1_ are taking around 9 minutes with the NVIDIA GeForce GTX 285 and with the allowed concurrent blocks set to 40000. The GPU isn't connected to the display so no lag is witnessed by me. Edit: It appears to be returning invalid results therefore it has been temporarily deprecated.
8) Message boards : Number crunching : GPU Requirements [OLD] (Message 35846) Posted 19 Jan 2010 by Anthony Waters Post: Milkyway@Home requires a GPU supporting Double Precision arithmetic. NVIDIA: Compute Capability 1.3 and Above Example Products: Geforce GTX 295 Geforce GTX 285 Geforce GTX 280 Geforce GTX 275 (credits to Bruce) Geforce GTX 260 Tesla S1070 Tesla C1060 Quadro Plex 2200 D2 Quadro FX 5800 Quadro FX 4800 (Based on GT200 GPU) ATI: Example Products: ATI HD Radeon 5970 (credits to kashi) ATI HD Radeon 5870 ATI HD Radeon 5850 ATI HD Radeon 4890 ATI HD Radeon 4870 ATI HD Radeon 4850 ATI HD Radeon 4770 ATI HD Radeon 4830 ATI HD Radeon 38x0 (credits to cenit for the AMD documentation describing the products above) ATI Firestream 9270 ATI Firestream 9250 ATI Firestream 9170 (credits to Cluster Physik) Please post other working products and this list will be updated as time goes on.
9) Message boards : News : ATI Application (Message 35614) Posted 13 Jan 2010 by Anthony Waters Post: After a few weeks of testing the deployment of the ATI application through BOINC it appears as most of the kinks have been worked out. Milkyway@Home would like to thank Cluster Physik for the development of the application and the many volunteers that assisted in reporting issues. Currently the application is available for Windows 32/64 bit and requires an ATI GPU with double precision capabilities. Milkyway@Home would also like to thank Advanced Micro Devices (AMD) for their generous donation of hardware.
10) Message boards : Application Code Discussion : Source Code of ATi app (Message 35577) Posted 12 Jan 2010 by Anthony Waters Post: The PTX code can be generated from the source, but I'm not well versed in how exp() would be calculated, what method was used to calculate it in the ATI version above?
11) Message boards : Number crunching : Testing ATI Application Availability (Message 35503) Posted 11 Jan 2010 by Anthony Waters Post: yes your correct it's a vista 32 bit with catalyst control 9.12 but what can i do to solve it or is it a server side issue, should i just go back to app_info for now? If you feel up to the task for some debugging can you: 1. look in C:\WINDOWS\system32 for the dll files aticalcl32.dll aticaldd32.dll aticalrt32.dll 2. In the BOINC application directory (C:\Documents and Settings\All Users\Application Data\BOINC for me), there should be a directory called 'slots' and within it subdirectories named 0,1,2,etc. Please verify that within the sub directories if there is a program named milkyway_0.21_windows_intelx86__ati13ati.exe that there is also a file named brook32.dll. If no sub directories are in the 'slots' folder please start a WU without the anonymous platform and once it starts immediately suspend activity, after that hopefully one of the 'slots' sub directories stays around.
12) Message boards : Number crunching : Testing ATI Application Availability (Message 35499) Posted 11 Jan 2010 by Anthony Waters Post: It executes correctly for me by manually downloading the files and manually executing the executable, does that error code mean the wrong dlls were found?
13) Message boards : Number crunching : Testing ATI Application Availability (Message 35497) Posted 11 Jan 2010 by Anthony Waters Post: Boinc just download the new ati13 ati apps but then immediately crashed. It managed to mangle several work units before i stopped it. Here's the error i got. Any ideas? running Boinc: 6.10.25 GPU: HD4850 http://milkyway.cs.rpi.edu/milkyway/result.php?resultid=27087886 <core_client_version>6.10.25</core_client_version> <![CDATA[ <message> - exit code -1073741757 (0xc0000043) </message> ]]> Thanks Mike P. What Catalyst Driver version and Operating System?
14) Message boards : Number crunching : Testing ATI Application Availability (Message 35486) Posted 11 Jan 2010 by Anthony Waters Post: The version on BOINC is actually 0.20b.
15) Message boards : Number crunching : Testing ATI Application Availability (Message 35470) Posted 11 Jan 2010 by Anthony Waters Post: The Windows 32 Bit version is now available on BOINC, please report any issues in this thread.
16) Message boards : Number crunching : Testing ATI Application Availability (Message 35210) Posted 8 Jan 2010 by Anthony Waters Post: I can confirm that Catalyst 9.12 works, I was able to complete several work units in a row, thanks to The Gas Giant for bringing this to my attention. Currently only the Windows 64 bit version will be available and I think it relies on Catalyst 9.7+ for Windows 7/Vista, and 9.12 for Windows XP users. If all goes well the Windows 32 bit version will also be deployed on BOINC.
17) Message boards : Number crunching : Testing ATI Application Availability (Message 35143) Posted 7 Jan 2010 by Anthony Waters Post: The signature verification error was resolved, it was an actual GPU crash. The result's of the WU that caused the crash were not uploaded to BOINC.
18) Message boards : Number crunching : Testing ATI Application Availability (Message 35140) Posted 7 Jan 2010 by Anthony Waters Post: The application was available for a short time but it kept causing GPU crashes with my configuration, therefore it was deprecated.
19) Message boards : Number crunching : Testing ATI Application Availability (Message 35119) Posted 6 Jan 2010 by Anthony Waters Post: Over the next 48 hours the Windows 64 Bit version of the ATI application will be available on BOINC. Since the ATI application has more dependencies on different dll files and appears to have some compatibility issues with different Catalyst driver versions and Windows versions it is expected that there there will be some issues occurring. Please have patience during this testing period and report any issues with the deployment in this thread. When reporting issues please include the Windows version, Catalyst driver version, and ATI graphics chipset. If there are a limited amount of issues the 32 bit version will follow.
20) Message boards : Number crunching : Linux x64 CUDA app - HOWTO?.. (Message 34683) Posted 19 Dec 2009 by Anthony Waters Post: Is it acceptable to assume that it worked prior to v0.24?

Next 20