How can you tell if CUDA is working? |
![]() |
| log in |
Questions and Answers : GPU applications : How can you tell if CUDA is working?
| Author | Message |
|---|---|
|
I've read many posts, and it seems people can see what the GPU is running, I can't. I have an i7 system, with a 9800GT. BOINC shows 8 tasks running. I am running 6.6.20 and Vista 64. I have edited my xml to run AP and MB, the lunatics SSE3 for AP and SSE4.1 for MB. The CPU is running both AP & MB. What else do I need to do? Thanks. | |
| ID: 887272 · | |
|
If you are using optimized apps via an app_info.xml file, then you are telling BOINC that only the applications listed in the app_info.xml file are available to run. Since the optimized AP and MB apps are not CUDA capable, you would need to locate either the stock app, download it an add it to your app_info.xml file, or locate the Raistmer optimized app, download it and add it to your app_info.xml file. | |
| ID: 887275 · | |
|
Thank you for that response. I DL'ed the Raistmer package and added it to the SETI folder. Now my tasks screen is very different. It is running: setiathome_enhanced 6.08 (cuda), and the status is: Running (0.04 CPUs, 1 CUDA) | |
| ID: 887282 · | |
|
Yes, the .04 CPUs is how much CPU power is needed to load the app into the GPU and save the result file when done. | |
| ID: 887287 · | |
|
I DL'ed the: Raistmer's_opt_package_V7_VLAR_kill.rar, there was that one, and Raistmer's_opt_package_V7.rar. Can I just insert my old xml file into the new one and run AP + MB on my CPU's? | |
| ID: 887292 · | |
|
No you may want to use raistmers v.10 app. the app_info may need to be editted. I can't remember if he included MB for the CPU on the app_info. Otherwise just install and restart and run. | |
| ID: 887303 · | |
|
Perhaps it's cheaper to read other threads, such as this one in Number Crunching and this one also in Number Crunching. Why try to reinvent the wheel when someone has done it before you? ;-) | |
| ID: 887322 · | |
Questions and Answers : GPU applications : How can you tell if CUDA is working?
| Copyright © 2013 University of California |