Not Getting WU

[bwm]

Originally Posted by J_M

Others on the Stanford site are reporting the same. It seems you should consider yourself lucky to even get that as the Nvidia WUs are the only ones going out consistently.

All of our GPU units here are nVidia and they have ALL been waiting for work more than 6 hours now. None of my SMP clients are having any problems at all...... yet.

[Dr.CSM]

Originally Posted by J_M

I wish Pande Group would be a bit more respectful of what we do and take the time to keep us informed instead of providing either no info or BS spin when there are problems.

Excellent point. Stanford doesn't seem to realize what kind of money people are spending on doing their work, especially now. And this is the return for it?? I'll bet that more than a few leave, including old-timers, never to return. They beter be thinking about it very hard.

[bwm]

Good news - GPU servers appear to be working again at about midnight Mountain time. All 9 of the completed GPU jobs I had in the queue have uploaded, fetched new ones, and rolling right along I still see some evidense of SMP being behind the curve so kep an eye on that. I think it will be quite some time until GPU stats are caught up though, due to the backload of GPU uploads very likely pending. We may next see a rash of can't get fresh GPU job errors.

[3.0charlie]

Originally Posted by chriskwarren

Interesting. My bigadv folding rig (using Linuxrouter's vm) that has been folding bigadvs for ages just picked up a regular wu (project 2662...1920pts according to HFM).

This vm has not been reset or anything.....

Normal. If no bigadv WUs are available, then the server will send an advmethods WU. And if those are not available a regular WU will be sent.

Don't know if the outage is related, but I lost a bigadv WU at 22:00 , 100% completion last night... can't see the log, says it's corrupted.

[geokilla]

Originally Posted by 3.0charlie

Normal. If no bigadv WUs are available, then the server will send an advmethods WU. And if those are not available a regular WU will be sent.

Don't know if the outage is related, but I lost a bigadv WU at 22:00 , 100% completion last night... can't see the log, says it's corrupted.

Ouch. Do you have the backup script set up? I'm trying to set it up but stuck at this step...

[Dr.CSM]

Originally Posted by bwm

Good news - GPU servers appear to be working again at about midnight Mountain time. All 9 of the completed GPU jobs I had in the queue have uploaded, fetched new ones, and rolling right along I still see some evidense of SMP being behind the curve so kep an eye on that. I think it will be quite some time until GPU stats are caught up though, due to the backload of GPU uploads very likely pending. We may next see a rash of can't get fresh GPU job errors.

Forget about that potential problem. Never materialized. Even regular -advmethods workunits are flowing in and out smoothly.

Now if I can figure out what you guys are talking about with this "big advanced" or whatever it is, maybe hubby and I can squeeze a bit more out of these ancient quads and GPU boxes.

[J_M]

Originally Posted by Dr.CSM

Now if I can figure out what you guys are talking about with this "big advanced" or whatever it is, maybe hubby and I can squeeze a bit more out of these ancient quads and GPU boxes.

-bigadv WUs award bonus points (worth 55k +) but they require a fast i7 with lots of memory to process.

[lowfat]

Are people still having a problem w/ getting a bigadv?

[3.0charlie]

Picked-up one finally yesterday afternoon.

[JayDeeHobbit]

I'm getting the "attempting to get work packet" message for the nvidia GPU console client. Are others still getting it?

The reason I'm wondering is because I just installed Win7 last night and I'm not sure if I am using the wrong client or if it just Stanford.