Message boards :
Graphics cards (GPUs) :
New beta application 6.18 for Windows
Message board moderation
| Author | Message |
|---|---|
GDFSend message Joined: 14 Mar 07 Posts: 1958 Credit: 629,356 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() |
This one should work. gdf |
nenymSend message Joined: 31 Mar 09 Posts: 137 Credit: 1,431,087,071 RAC: 58,001 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
The first observation http://photo.uloziste.com/?fotka=ba59aa18c336fa66.jpg&d=3a43210e71c8c82b&size=0 ID 3129 XP x_64, GTX260-65nm, one of these unhappy ones crashing units until 185.xx drivers. 100% one core load, 31 min 12% progress. Link is to screenshot GPU-Z + CUDA-Z + task manager. |
|
Send message Joined: 11 Jul 09 Posts: 1639 Credit: 10,159,968,649 RAC: 2 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
Similar observation. A26-TONI_KKBETA2-0-100-RND8274_4 (I seem to be the fifth to try it) is running, but using 100% of a CPU core: after 40 minutes (5%), CPU time is only one second less than elapsed time. Because BOINC sets GPU 'feeder' tasks at a higher priority than pure CPU tasks, it's stealing time from all four of the other BOINC tasks on this quad. |
|
Send message Joined: 18 Aug 08 Posts: 8 Credit: 127,707,074 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
|
|
Send message Joined: 17 Aug 08 Posts: 2705 Credit: 1,311,122,549 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
What about its GPU performance? MrS Scanning for our furry friends since Jan 2002 |
nenymSend message Joined: 31 Mar 09 Posts: 137 Credit: 1,431,087,071 RAC: 58,001 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
"nenym" wrote: ID 3129 XP x_64, GTX260-65nm, one of these unhappy ones crashing units until 185.xx drivers. Two task crashed, the first after 6,425s (GPU little OCed), second after 3,762s (factory clock settings). Hmm...that weird 65nm GPU is still unusable for GPUGRD. Switching to Seti. |
AardvarkSend message Joined: 27 Nov 08 Posts: 28 Credit: 82,362,324 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
Looks good 12 Mar 2010 16:23:54 UTC 12 Mar 2010 22:36:47 UTC Completed and validated 17,086.48 16,753.04 4,535.61 6,803.41 ACEMD beta version v6.18 (cuda). as compaired to (typical values) 1 Mar 2010 5:48:44 UTC 1 Mar 2010 18:41:37 UTC Completed and validated 23,123.93 3,017.39 4,535.61 6,123.07 ACEMD beta version v6.10 (cuda) Uses 1 multithread CPU core per GPU ( 2X GTX260, 55nm, XFX factory OC). That's about 26% faster...... Boinc CPU tasks in progress at same time. (8 tasks shared over 6 threads). |
X-Files 27Send message Joined: 11 Oct 08 Posts: 95 Credit: 68,023,693 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
Not good. It occupies 1 core.
|
|
Send message Joined: 11 Jul 09 Posts: 1639 Credit: 10,159,968,649 RAC: 2 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
What about its GPU performance? Not good on a 9800GT. Maybe 5% faster than v6.03, but at the expense of taking an entire CPU core - over 40K seconds taken away from other projects. But at least it finished and validated. |
GDFSend message Joined: 14 Mar 07 Posts: 1958 Credit: 629,356 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() |
The 100% CPU core is a not intended and will be removed. Don't know why is there. This application is 30-40% faster on newer cards only based on G200 chip. For older cards, it will be the same speed of the current main application. gdf |
|
Send message Joined: 24 Dec 08 Posts: 738 Credit: 200,909,904 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Got this one. Its certainly was quicker but as the other guys have said uses a heck of a lot of CPU to get there. Was run on a GTX295. I notice its got lots more device info in the messages, but still doesn't say which device its running on. BOINC blog |
GDFSend message Joined: 14 Mar 07 Posts: 1958 Credit: 629,356 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() |
No the CPU usage is just unrelated. gdf |
|
Send message Joined: 17 Aug 08 Posts: 2705 Credit: 1,311,122,549 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
This application is 30-40% faster on newer cards only based on G200 chip. Wow 30-40% faster on G200+ is just great! This should also help the GT240, which is already quite popular for GPU-Grid. Seems like the enhancements in the new architecture are finally paying off. MrS Scanning for our furry friends since Jan 2002 |
|
Send message Joined: 18 Aug 08 Posts: 8 Credit: 127,707,074 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
|
|
Send message Joined: 4 Apr 09 Posts: 450 Credit: 539,316,349 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
GDF ... you guys are GOOD !!! I just processed my first beta from this round on a dual card GTX295 in 3.25 hours. I see now why you are talking about increasing WU size :-) Thanks - Steve |
|
Send message Joined: 12 Feb 09 Posts: 57 Credit: 23,376,686 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
http://www.gpugrid.net/result.php?resultid=1984108 my first beta WU, ran successfully and with a nice speedup, as reported before, the cpu usage was for me 100% too. |
skgivenSend message Joined: 23 Apr 09 Posts: 3968 Credit: 1,995,359,260 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
25% complete for a Beta 6.18 on a GTX260sp216 (Phenom II 940) and it already looks much faster. Estimated time to finish is 4h 30min. The last 6.03 task on same system took 5h 50min. The 6.03 prior to that was 5h 56min.101. Did not get any on my GT240's. |
skgivenSend message Joined: 23 Apr 09 Posts: 3968 Credit: 1,995,359,260 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Took about 33% less time to complete one 6.18 task than a 6.03 task, albeit at the expense of CPU time (required a full CPU). Again, I am quite happy with this as GPU work should be prioritised in whatever way required, but others might go on about it. Thanks, |
|
Send message Joined: 28 Mar 09 Posts: 490 Credit: 11,739,145,728 RAC: 95,752 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
I downloaded 2 of these units. They each took a little more than 3 hours to complete with a GTX 285 card, as oppose to over 4 hours for the ACEMD 6.03. The only downside is that it also uses up 100% of one my CPU's time. Over all, this is very good. |
©2026 Universitat Pompeu Fabra