Bug in server affecting older BOINC clients with NVIDIA GPUs.

Message boards : News : Bug in server affecting older BOINC clients with NVIDIA GPUs.

Author	Message
Eric J Korpela Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 15 Mar 05 Posts: 953 Credit: 6,139,304 RAC: 8,290	Message 43621 - Posted: 17 Aug 2012, 5:38:16 UTC
	We've identified a bug in the current BOINC server that is online at SETI@home. With older BOINC clients this bug results in running multiple SETI@home GPU applications simultaneously on a single GPU. While we debug and fix the problem we've suspended distribution of NVIDIA work. We hope that everything will be back to normal some time tomorrow. ____________
	ID: 43621 ·

Richard Haselgrove Volunteer tester Send message Joined: 3 Jan 07 Posts: 969 Credit: 2,070,629 RAC: 18	Message 43624 - Posted: 17 Aug 2012, 9:47:23 UTC - in response to Message 43621.
	So far as I can tell, both the bug itself and the temporary suspension (by deprecating the stock apps) seems to apply to stock only - Anonymous Platform appears to be running normally.
	ID: 43624 ·

Richard Haselgrove Volunteer tester Send message Joined: 3 Jan 07 Posts: 969 Credit: 2,070,629 RAC: 18	Message 43625 - Posted: 17 Aug 2012, 11:48:22 UTC
	David has coded a provisional fix in [26042]. The fix is in writing the app_version data to the sched_reply file: the full app_version data isn't sent to anonymous hosts, just the <platform>, <version_num> and <plan_class> for each result, so BOINC can pick the right one out of our app_info files. Our sched_request files are already properly tested for both CUDA and NVIDIA strings on receipt, so this fix should work.
	ID: 43625 ·

Sten-Arne Volunteer tester Send message Joined: 10 Mar 12 Posts: 484 Credit: 2,544,835 RAC: 6,926	Message 43627 - Posted: 17 Aug 2012, 15:00:20 UTC - in response to Message 43621.
	We've identified a bug in the current BOINC server that is online at SETI@home. With older BOINC clients this bug results in running multiple SETI@home GPU applications simultaneously on a single GPU. While we debug and fix the problem we've suspended distribution of NVIDIA work. We hope that everything will be back to normal some time tomorrow. Hmm, so it was a bug then, and not something wrong with my two Nvidia computers?
	ID: 43627 ·

Eric J Korpela Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 15 Mar 05 Posts: 953 Credit: 6,139,304 RAC: 8,290	Message 43628 - Posted: 17 Aug 2012, 16:01:29 UTC - in response to Message 43627.
	Yes ____________
	ID: 43628 ·

Sten-Arne Volunteer tester Send message Joined: 10 Mar 12 Posts: 484 Credit: 2,544,835 RAC: 6,926	Message 43629 - Posted: 17 Aug 2012, 18:38:09 UTC - in response to Message 43628. Last modified: 17 Aug 2012, 18:50:21 UTC
	Yes Thanks. The bug fix seems to work. My Samsung is now downloading new 6.04's, at the moment 2 down, and only one running. 0.04CPU +1GPU. Same as before the problem showed itself. Now, I'll make the ION download a few 6.04's, to see if it works with that one too. Edit, Added: ION now have two 6.04's downloaded, and only one is running. 0.11CPU +1GPU. Same as before the problem showed itself. So, at least for my computers, the issue seems resolved.
	ID: 43629 ·

Eric J Korpela Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 15 Mar 05 Posts: 953 Credit: 6,139,304 RAC: 8,290	Message 43642 - Posted: 21 Aug 2012, 2:05:18 UTC - in response to Message 43629.
	Yay ____________
	ID: 43642 ·

Message boards : News : Bug in server affecting older BOINC clients with NVIDIA GPUs.

AstroPulse is funded in part by the NSF through grant AST-0307956