Message boards :
Number crunching :
Errors using Intel GPU
Message board moderation
Author | Message |
---|---|
James W Send message Joined: 26 May 12 Posts: 51 Credit: 4,956,027 RAC: 13 |
I recently started noticing errors on tasks crunched with Intel GPU, specifically WU 1689137533. My results on task 3937904699 were exactly the same as my wingman's, though I errored out and he is pending validation. We're both using Intel GPU, though he's using the Anonymous platform. My Stderr message was "finish file present too long." I happened to be using my BOINC Manager in a similar situation where I noted my GPU task was sitting at 100% in active tasks for some time, until I updated SETI work. Not a big issue at this point, but don't want to see this being an ongoing issue either. What peeked my concern was the fact that my job errored out and wingman's didn't (as of yet). Any ideas on this? Thanks. |
MarkJ Send message Joined: 17 Feb 08 Posts: 1139 Credit: 80,854,192 RAC: 5 |
Any ideas on this? Thanks. The most common issue with the Intel HD graphics is their drivers. Basically only 1 or 2 of them work properly for crunching. I don't have HD2500, but I do have machines with HD4000 Graphics and use the 10.18.10.3621 drivers (Win7 64 bit). I would suggest you try and get that version if you can. Intel have a habit of hiding the older drivers though so you might not be able to find them in the Intel Download Center. Richard might have a link if you can't find it. There are a few message threads going about the same issue but the users concerned might be using different models of Intel GPU's, such as this thread BOINC blog |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
In the Intel gpu not seen by BOINC thread we talked about driver a lot. Richard did some testing with his Ivy Bridge & Haswell hardware with several drivers. HD 4000 with drivers 10.18.10.3621 and 10.18.10.4061 HD 4600 with drivers 10.18.10.3621 and 10.18.14.4080 In those tests the only one that didn't come out well was the 10.18.14.4080 driver. So either of the other two should be good. 10.18.10.4061 is a newer driver that hasn't shown to have any issues thus far, but 10.18.10.3621 is the previous known good driver. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
I recently started noticing errors on tasks crunched with Intel GPU, specifically WU 1689137533. My results on task 3937904699 were exactly the same as my wingman's, though I errored out and he is pending validation. We're both using Intel GPU, though he's using the Anonymous platform. My Stderr message was "finish file present too long." I happened to be using my BOINC Manager in a similar situation where I noted my GPU task was sitting at 100% in active tasks for some time, until I updated SETI work. This should not be a GPU driver issue. Once crunching is done, the app calls a boinc_finish routine which creates the finish file and should go on to self-immolation. IOW, it's all related to CPU code supplied by the BOINC developers, both the part of the API which must be built into applications and the BOINC client itself. Joe |
©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.