Message boards :
Number crunching :
Optimize your GPU. Find the value the easy way.
Message board moderation
Previous · 1 . . . 3 · 4 · 5 · 6 · 7 · 8 · 9 . . . 13 · Next
Author | Message |
---|---|
S@NL - eFMer - efmer.com/boinc Send message Joined: 7 Jun 99 Posts: 512 Credit: 148,746,305 RAC: 0 |
I agree the WU is a bit short, on some of my systems it takes 5 seconds to load. But it's a nice starting point and most will be hesitant to do a test that takes a couple of hours. But I plan to add a couple more WU to test. TThrottle Control your temperatures. BoincTasks The best way to view BOINC. Anza Borrego Desert hiking. |
Horacio Send message Joined: 14 Jan 00 Posts: 536 Credit: 75,967,266 RAC: 0 |
Ive made that just adding another set (folder) with the "long" WU, so in that way I tested first automatically with the quick one and only used the long one for manual tests when the results were very similar between different counts... thats very easy, no need to add any code, just another folder... |
Cherokee150 Send message Joined: 11 Nov 99 Posts: 192 Credit: 58,513,758 RAC: 74 |
Thanks for the great performance evaluator app! By the way, it seems to work great on NVIDIA driver version 266.58, too! It appears I should run probably two or, perhaps, three units on my GPU for best performance (three will depend on how much it affects my CPU units); at least that's how the following excerpted results look to me: Results: Device: 0, device count: 1, average time / count: 173, average time on device: 173 Seconds (2 Minutes, 53 Seconds) Device: 0, device count: 2, average time / count: 313, average time on device: 156 Seconds (2 Minutes, 36 Seconds) Device: 0, device count: 3, average time / count: 460, average time on device: 153 Seconds (2 Minutes, 33 Seconds) Currently I'm running the following (only pertinent info here): Starting BOINC client version 7.0.25 for windows_x86_64 Processor: 4 GenuineIntel Intel(R) Core(TM)2 Quad CPU Q8400 @ 2.66GHz [Family 6 Model 23 Stepping 10] Microsoft Windows Vista: Ultimate x64 Edition, Service Pack 2, (06.00.6002.00) NVIDIA GPU 0: GeForce GTX 460 (driver version 266.58, CUDA version 3.20, compute capability 2.1, 993MB, 753MB available, 1164 GFLOPS peak) OpenCL: NVIDIA GPU 0: GeForce GTX 460 (driver version 266.58, device version OpenCL 1.0 CUDA, 993MB, 753MB available) The box is overclocked. I've got it running -very- cool and stable with CPU at 3.04GHz Core, 1520 MHz FSB, and 760 Mhz DRAM; GPU at 866 MHz Core, 1732 MHz shaders and 2160 MHz Memory. I haven't done anything with SETI optimization, however, so I'm afraid I'm a novice in this area. I've been looking through all the posts, and I must admit I'm totally confused about how to get my system to run more than one GPU unit at a time. Here's my question: How do I set up SETI and code a config.xml file to multiple process on the GPU? Also, can I do it with the stock SETI apps, or do I need to run a special lunatics version? I've been in this almost since the beginning and have always pushed my boxes to the max through much personal tweaking and prodding. I do this because I truly believe in helping the advancement of science and the benefit of mankind. So, can you please help me? Once I learn how to use your optimization for my "main man" (the Quad listed above), I can easily set up the others. Thanks so much for any assistance you can provide. Finally, although I don't post often, I have been following many of the SETI threads for years, so I would like to take this opportunity to say: HATS OFF TO ALL OF YOU!!! |
S@NL - eFMer - efmer.com/boinc Send message Joined: 7 Jun 99 Posts: 512 Credit: 148,746,305 RAC: 0 |
V 1.2 With some small bug fixes and brand new graphs. http://www.efmer.eu/forum_tt/index.php?topic=974.msg5946 TThrottle Control your temperatures. BoincTasks The best way to view BOINC. Anza Borrego Desert hiking. |
Cherokee150 Send message Joined: 11 Nov 99 Posts: 192 Credit: 58,513,758 RAC: 74 |
Thanks Fred. You do reply fast! (I thought I might have to wait a day or so, you know.) I'll muddle through this new info and do some old-fashioned trial and error. If I run into any problems, I'll let you know. Thanks again. :) |
JohnDK Send message Joined: 28 May 00 Posts: 1222 Credit: 451,243,443 RAC: 1,127 |
V 2.0 With some small bug fixes and brand new graphs. Have just downloaded it and the app window says v1.2 ? |
S@NL - eFMer - efmer.com/boinc Send message Joined: 7 Jun 99 Posts: 512 Credit: 148,746,305 RAC: 0 |
V 2.0 With some small bug fixes and brand new graphs. Fixed, should be 1.2. TThrottle Control your temperatures. BoincTasks The best way to view BOINC. Anza Borrego Desert hiking. |
S@NL - eFMer - efmer.com/boinc Send message Joined: 7 Jun 99 Posts: 512 Credit: 148,746,305 RAC: 0 |
Install this one: http://lunatics.kwsn.net/index.php?module=Downloads;catd=9 Make sure you have the minimum driver mentioned in the installer and choose the bit version. Next go here: C:\ProgramData\BOINC\projects\setiathome.berkeley.edu Open app_info.xml with notepad.exe. Look for these 2 lines together: <type>CUDA</type> <count>1</count> Change the 1 in count into 0.33 or 0.5 Make sure you use a dot . after the 0! TThrottle Control your temperatures. BoincTasks The best way to view BOINC. Anza Borrego Desert hiking. |
Cherokee150 Send message Joined: 11 Nov 99 Posts: 192 Credit: 58,513,758 RAC: 74 |
Thank you again, Fred! This is what I need to know. I'll back up BOINC and test it out. And don't worry, I'll be sure to include the decimal point. Wouldn't that be fun if I put in 31!! lol ;) |
S@NL - eFMer - efmer.com/boinc Send message Joined: 7 Jun 99 Posts: 512 Credit: 148,746,305 RAC: 0 |
Thank you again, Fred! Other countries use a , The . isn't that common..... TThrottle Control your temperatures. BoincTasks The best way to view BOINC. Anza Borrego Desert hiking. |
Cherokee150 Send message Joined: 11 Nov 99 Posts: 192 Credit: 58,513,758 RAC: 74 |
Hi Fred, I love the improvements you've made to ver 1.2. I did notice a slight modification you might need to make to your new graph. The 30 minute option has a slight flaw that causes the high and low spikes to jump/change their peaks every other update, i.e., they display one set of peaks and valleys across the 30 minute plot, then the peaks and valleys change height in the next update. On the third update, they revert to the same as the first. The fourth reverts to the second, etc. This appears to occur -only- on the 30 minute option. It's a minor change, I suspect, but your program is so good I thought you'd probably want to take a look at it. Thank you again! :) |
S@NL - eFMer - efmer.com/boinc Send message Joined: 7 Jun 99 Posts: 512 Credit: 148,746,305 RAC: 0 |
Maybe my mathematical skills. The problem is there are too many points, so some have to be condensed. Normally you would have 50,60,70,80,50,70 Condensed [50,60],[70,80],[50,70] 55, 75, 60 One second later adds a new one 90, [90,50],[60,70],[80,50] 70, 65, 65. That are entirely different values. One option would be to update every 2 seconds in this case, so that everything would align again. TThrottle Control your temperatures. BoincTasks The best way to view BOINC. Anza Borrego Desert hiking. |
Cherokee150 Send message Joined: 11 Nov 99 Posts: 192 Credit: 58,513,758 RAC: 74 |
Oh, for sure, I understand, Fred. I've traveled quite extensively, so I'm a bit more "in the know" than most people from the U.S. It sure would stir up the pot a bit if I put in a "," or left it out completely! That's why I said, tongue-in-cheek, as perhaps a little pun, "wouldn't it be fun, if I put in thirty-one!" lol ;) Well, the sun's been up for a bit, so I suppose I better get a little "shut-eye" before I get busy with the chores. Guess my GPU fun must wait 'til I'm done. CUL (old amateur radio/morse code term, actually) If you know code, then you'll understand -.-. .... --- .-- (hint: it's Italian) |
S@NL - eFMer - efmer.com/boinc Send message Joined: 7 Jun 99 Posts: 512 Credit: 148,746,305 RAC: 0 |
Like hungry. TThrottle Control your temperatures. BoincTasks The best way to view BOINC. Anza Borrego Desert hiking. |
Cherokee150 Send message Joined: 11 Nov 99 Posts: 192 Credit: 58,513,758 RAC: 74 |
One option would be to update every 2 seconds in this case, so that everything would align again. That would probably make it read easier. It's just a bit disconcerting watching the already posted graph values appear to change every other second. As I suspected, only a minor change. A bientôt! |
Cherokee150 Send message Joined: 11 Nov 99 Posts: 192 Credit: 58,513,758 RAC: 74 |
I figured you'd know. Yes, in fact, I could use some right now. Oh, gotta go...time to awaken my son! пока (Russian) |
John Thon Send message Joined: 12 May 08 Posts: 18 Credit: 12,310,310 RAC: 34 |
Here are the results for my GTX460, What do I do with it or is it just to show what my video card can do? Running Windows 7 Starting automatic test: (x41g) 01 September 2012 - 16:06:02 Start, devices: 1, device count: 1 (1.00) 01 September 2012 - 16:10:39 Runtime: Device: 0, count: 0, 276 seconds 01 September 2012 - 16:10:39 Device: 0, Count: 0, finished. Ready --------------------------------------------------------------------- Results: Device: 0, device count: 1, average time / count: 276, average time on device: 276 Seconds (4 Minutes, 36 Seconds) Next --------------------------------------------------------------------- 01 September 2012 - 16:10:40 Start, devices: 1, device count: 2 (0.50) 01 September 2012 - 16:17:05 Runtime: Device: 0, count: 1, 383 seconds 01 September 2012 - 16:17:05 Device: 0, Count: 1, finished. 01 September 2012 - 16:17:06 Runtime: Device: 0, count: 0, 384 seconds 01 September 2012 - 16:17:06 Device: 0, Count: 0, finished. Ready --------------------------------------------------------------------- Results: Device: 0, device count: 2, average time / count: 383, average time on device: 191 Seconds (3 Minutes, 11 Seconds) Next --------------------------------------------------------------------- 01 September 2012 - 16:17:07 Start, devices: 1, device count: 3 (0.33) 01 September 2012 - 16:26:10 Runtime: Device: 0, count: 0, 542 seconds 01 September 2012 - 16:26:10 Device: 0, Count: 0, finished. 01 September 2012 - 16:26:16 Runtime: Device: 0, count: 2, 547 seconds 01 September 2012 - 16:26:16 Device: 0, Count: 2, finished. 01 September 2012 - 16:26:18 Runtime: Device: 0, count: 1, 550 seconds 01 September 2012 - 16:26:18 Device: 0, Count: 1, finished. Ready --------------------------------------------------------------------- Results: Device: 0, device count: 3, average time / count: 546, average time on device: 182 Seconds (3 Minutes, 2 Seconds) Next --------------------------------------------------------------------- 01 September 2012 - 16:26:18 Start, devices: 1, device count: 4 (0.25) 01 September 2012 - 16:38:28 Runtime: Device: 0, count: 3, 729 seconds 01 September 2012 - 16:38:28 Runtime: Device: 0, count: 0, 729 seconds 01 September 2012 - 16:38:28 Runtime: Device: 0, count: 2, 729 seconds 01 September 2012 - 16:38:28 Device: 0, Count: 0, finished. 01 September 2012 - 16:38:28 Device: 0, Count: 3, finished. 01 September 2012 - 16:38:28 Device: 0, Count: 2, finished. 01 September 2012 - 16:38:30 Runtime: Device: 0, count: 1, 731 seconds 01 September 2012 - 16:38:30 Device: 0, Count: 1, finished. Ready --------------------------------------------------------------------- Results: Device: 0, device count: 4, average time / count: 729, average time on device: 182 Seconds (3 Minutes, 2 Seconds) The best average time found: 182 Seconds (3 Minutes, 2 Seconds), with count: 0.33 (3) Thanks, John |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13722 Credit: 208,696,464 RAC: 304 |
or is it just to show what my video card can do? Yep. Grant Darwin NT |
Cherokee150 Send message Joined: 11 Nov 99 Posts: 192 Credit: 58,513,758 RAC: 74 |
Hi John! Actually, this tool can be used for more than just showing you how fast your GPU is. In your case, it shows that you might benefit by running two, and perhaps even three, SETI units simultaneously on your GPU. The average run time of the multiple units compared to the initial baseline of just one unit (the way it normally runs) gives you a good idea of how much faster your units will process this way. Your GPU could possibly complete units nearly 45% faster running two at a time, or almost 52% faster running three at a time. You just need to verify your computer's graphics performance for the rest of the system doesn't suffer as a result. Also, John, you should make sure the GPU doesn't run too hot. One other possible drawback is that your CPU units will lose some performance speed. You should check this to ensure the overall average rate of processing of all units doesn't go down. With two GPU units at a time, you will almost certainly benefit. Three may be somewhat marginally better overall, but, as each system is different, you will need to verify this yourself. Fred has provided good instructions and links to more information and to the software you can install to achieve this goal. If you're interested, John, check out the posts in this thread and you will find what you need. David :) |
MadMaC Send message Joined: 4 Apr 01 Posts: 201 Credit: 47,158,217 RAC: 0 |
What an absolutely brilliant bit of software - hats off to you my man, and thx for the effort :-) |
©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.