Message boards :
Number crunching :
Head Scratcher???
Message board moderation
Previous · 1 · 2
Author | Message |
---|---|
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
Rob, I'm not sure since I only have one gpu and can't test, but I'm talking about the one that is higher in the file: I think you're both right. AFAIK, Jason uses the old GetPrivateProfileString method designed for handling .INI files - so unless the [bus1slot0] KeyName is activated (by removing the semi-colon comment signifier from that line), the following values remain global. But you'd probably stand less chance of confusing yourself when you re-visit the file in six months' time if you used the global block for a value intended to be global. Because this file is designed to be read by the application, there is no need to restart BOINC after making changes - it'll be read automatically each time a new task starts, and the current values used for that task. BOINC knows nothing of this. |
rob smith Send message Joined: 7 Mar 03 Posts: 22202 Credit: 416,307,556 RAC: 380 |
Right, I've done a quick and dirty trial - if you don't "un-semi-colon" the ;[bus1slot0]line the priority is affected for both GPU (on my set-up anyway); if you do then only one of the cards is affected. I hope that clears that one up. Further it looks as if going from below normal to normal gives somewhere between 5 and 10% improvement. I'll let it settle for a few days and see what goes down, or hopefully up.... Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
j tramer Send message Joined: 6 Oct 03 Posts: 242 Credit: 5,412,368 RAC: 0 |
http://www.gpureview.com/show_cards.php will give you a side by side review of the two cards |
j tramer Send message Joined: 6 Oct 03 Posts: 242 Credit: 5,412,368 RAC: 0 |
one has a bigger bandwidth, and the other has more flops both are 128 bit |
Fred E. Send message Joined: 22 Jul 99 Posts: 768 Credit: 24,140,697 RAC: 0 |
Philhnnss, it didn't occur to me the other day, but the 450's should be able to crunch cuda 42 rather than the cuda 32 you are running. See screenshot in this post in another thread: http://setiathome.berkeley.edu/forum_thread.php?id=73155&postid=1460069 The driver you have is okay, no need to upgrade. If in doubt, check the Boinc event log startup messages - the cuda version supported should be in the first 20 lines. Rob, Richard: thanks for the clarification. I've just always used that first priority line. Another Fred Support SETI@home when you search the Web with GoodSearch or shop online with GoodShop. |
Philhnnss Send message Joined: 22 Feb 08 Posts: 63 Credit: 30,694,327 RAC: 162 |
OK, cool. Plugging in the features of both cards it show the 450's have shader clocks running at 1631. The 650's run at 978. Every other number is VERY heavy in favor of the 650's. So I guess for future refrance for SETI work I'll compare the shader clock speed as well. THANK YOU!!! |
Philhnnss Send message Joined: 22 Feb 08 Posts: 63 Credit: 30,694,327 RAC: 162 |
Good deal. I'll upgrade the XP machine's Lunitics and Cuda as well. I am one of the ones the good folks at Lunitic's warn us about. When they had the run in with Microsoft and took down the installers I pretty much forgot to look if they were back. So I had not updated in all this time!! |
rob smith Send message Joined: 7 Mar 03 Posts: 22202 Credit: 416,307,556 RAC: 380 |
The straight numbers do not tell the whole story, as there is a technology change between the 4xx and 6xx series that allows the 6xx to do far more work per core than the 4xx. Nvidia have a useful set of graphs on their site which give a good, visual, indication of the relative performance: Look at top graph Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
Michael W.F. Miles Send message Joined: 24 Mar 07 Posts: 268 Credit: 34,410,870 RAC: 0 |
Higher RAC Crunching V7 on GPU is just not the same as the old V6. The credit given is small compred to say your GPU running AP tasks . 1 AMD 1100T with 5 cores active to crunch V7 Multibeam on a M4N72-E mobo which is DDR2 ram using Hyper X DDR2 1066 I use 1 Zotac 460 GTX 768 to crunch nothing but Astropulse and I am averaging 20000 RAC a day. If I use my GPU to crunch multibeam I will receive a much lower RAC. Drivers as well I have noticed an increase in speed and performance. The latest Nidia driver 331.93 has solved playback studder while crunching and watching movies at the same time. Try overclocking your NB clock and HT link to 2.6 or better if you can with the ram your using it will be much better for speed thus feeding your GPU will be faster. Phenoms like high NB clock and HT link When I upped mine from 2 GHZ to 2.6 the system performed much better. Michael Miles Team New Mexico |
Philhnnss Send message Joined: 22 Feb 08 Posts: 63 Credit: 30,694,327 RAC: 162 |
Wanted to say "THANK YOU" to all that helped me!! The changes I made with setizero, the computer with the 2 450's, didn't really make any improvements to my RAC. But the changes you guys suggested with SetiOne has made a profound improvement. I know RAC isn't really a true comparison, but it is easy to track with the statistic tab. And using that tab my RAC is climbing at an almost vertical rate!!! I am still going to hold off a couple of weeks to change the value from below normal to above normal. I want to see were it levels off at first. |
©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.