GPU Errors on Blanked WUs

留言板 : Number crunching : GPU Errors on Blanked WUs
留言板合理

To post messages, you must log in.

作者消息
IFRS
志愿者测试人员
Avatar

发送消息
已加入:21 May 99
贴子:1736
积分:259,180,282
近期平均积分:0
United States
消息 943420 - 发表于:28 Oct 2009, 20:34:13 UTC - 回复消息 943343.  

I just got a bunch of the new s/w processed WUs from '06; Reschedule sent them to my GPU. When they started to run, ALL of them (about 30) got computation errors in just a few seconds, and then several non-06 WUs got errors also. I shut down my machine, and restarted it, and so far (fingers crossed) have no more such errors.

Has anyone else had a GPU problem with these?

I am using an EVGA GTS 250 card, and the temp was only around 52 degrees, so not a heating problem.


Too little free GPU memory, you may need to reboot:
Total GPU memory 536870912 free GPU memory 31625216

                                                              Joe


That will explain a lot. Sometimes my GTX 295 just kill all my 3 day cache in minutes. A reboot each couple of days sounds a good measure until things cleared up.
ID: 943420 · 举报违规帖子
Cruncher-American Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor

发送消息
已加入:25 Mar 02
贴子:1513
积分:370,893,186
近期平均积分:340
United States
消息 943363 - 发表于:28 Oct 2009, 15:47:46 UTC

G and J:
Thanks for your responses.
I think that may have been it, as it seems to be working OK now that I rebooted.
But what could cause that problem all of a sudden? Is it just a random event?I've been running this card 24/7 for months now, and never seen this kind of phenomenon before...
ID: 943363 · 举报违规帖子
Profile Gundolf Jahn

发送消息
已加入:19 Sep 00
贴子:3184
积分:446,358
近期平均积分:0
Germany
消息 943348 - 发表于:28 Oct 2009, 15:09:19 UTC - 回复消息 943343.  

Too little free GPU memory, you may need to reboot:
Total GPU memory 536870912 free GPU memory 31625216

And if one task fails with that problem, all subsequent tasks will fail too until reboot.

Gruß,
Gundolf
ID: 943348 · 举报违规帖子
Josef W. Segur
志愿者开发人员
志愿者测试人员

发送消息
已加入:30 Oct 99
贴子:4504
积分:1,414,761
近期平均积分:0
United States
消息 943343 - 发表于:28 Oct 2009, 14:43:30 UTC - 回复消息 943332.  

I just got a bunch of the new s/w processed WUs from '06; Reschedule sent them to my GPU. When they started to run, ALL of them (about 30) got computation errors in just a few seconds, and then several non-06 WUs got errors also. I shut down my machine, and restarted it, and so far (fingers crossed) have no more such errors.

Has anyone else had a GPU problem with these?

I am using an EVGA GTS 250 card, and the temp was only around 52 degrees, so not a heating problem.


Too little free GPU memory, you may need to reboot:
Total GPU memory 536870912 free GPU memory 31625216

                                                              Joe
ID: 943343 · 举报违规帖子
Profile perryjay
志愿者测试人员
Avatar

发送消息
已加入:20 Aug 02
贴子:3377
积分:20,676,751
近期平均积分:0
United States
消息 943338 - 发表于:28 Oct 2009, 14:20:20 UTC - 回复消息 943332.  

I don't keep a large cache so I haven't had too many of the new WUs but the few I've done so far have run fine on both my CPU and my 9500GT GPU. I'm also running the rescheduler but only to move VLARs off my GPU so I can't help you there.

I think I've seen the problem you describe before but can't remember where. It was some time back and way before the new SW blanked WUs. I'm sure somebody will chime in with the link to it soon.


PROUD MEMBER OF Team Starfire World BOINC
ID: 943338 · 举报违规帖子
Cruncher-American Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor

发送消息
已加入:25 Mar 02
贴子:1513
积分:370,893,186
近期平均积分:340
United States
消息 943332 - 发表于:28 Oct 2009, 13:49:18 UTC

I just got a bunch of the new s/w processed WUs from '06; Reschedule sent them to my GPU. When they started to run, ALL of them (about 30) got computation errors in just a few seconds, and then several non-06 WUs got errors also. I shut down my machine, and restarted it, and so far (fingers crossed) have no more such errors.

Has anyone else had a GPU problem with these?

I am using an EVGA GTS 250 card, and the temp was only around 52 degrees, so not a heating problem.
ID: 943332 · 举报违规帖子

留言板 : Number crunching : GPU Errors on Blanked WUs


 
©2020 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.