Empty work server assignment:

[BluePilot]

Checked my folding rig this morning (UK time) to discover the GPU's sitting idle. Didn't mind too much as I had been trying to find an opportunity to upgrade the machine so took the chance to throw in another couple of GPU's which arrived a few days ago. The only issue is that I now have 4 GTX980's sitting idle! Even the Xeon CPU is running slower as I have had to steal a couple of cores to feed the new GPU's.

Hopefully there will be more Core17 WU's available soon.

BP

[billford]

Anyone know if there's any activity on this?

[jimerickson]

8 gtx980's sitting idle here. what a waste of resources. hope there are more work units soon.

[billford]

Last time I looked at serverstats there were over 200 WUs (it won't load for me atm), but the AS's seem to have taken a dislike to Maxwells running on Linux

[Buck Nasty]

No responses from PG either here or on Reddit. Hope this does not go unanswered beyond the close of business @ Stanford(PST).

[billford]

So do I, but I'm not going to hold my breath.

[7im]

Anyone that commented about getting low points on core 18 should have considered that getting no work units was worth even less points.

[ChristianVirtual]

My 780 got a new 9201 this morning !

Q1: are new WU generated
Q2: does the AS dislike Maxwell complete ?

WU received: 9201 (610,3,162)
gpu:1:GK110 [GeForce GTX 780]

[billford]

I think it's just Maxwell, my 780Ti got a P9201 earlier this evening, uploaded it a couple of hours ago.

Back to the usual P1300x now

[bruce]

There still are a number of unknowns about getting full support for Maxwell, but in this case, something else was fixed. We're back where we were the last time it worked. (Yesterday?)

[billford]

We're back where we were the last time it worked. (Yesterday?)

I'm not:

Code: Select all

00:26:36:FS01:Paused
00:26:42:FS01:Unpaused
00:26:58:WU00:FS01:Connecting to 171.67.108.200:80
00:26:59:WARNING:WU00:FS01:Failed to get assignment from '171.67.108.200:80': Empty work server assignment
00:26:59:WU00:FS01:Connecting to 171.67.108.204:80
00:26:59:WARNING:WU00:FS01:Failed to get assignment from '171.67.108.204:80': Empty work server assignment
00:26:59:ERROR:WU00:FS01:Exception: Could not get an assignment

edit- that's the Maxwell/Linux client.

[jimerickson]

still getting empty work server assignment here too.

[billford]

Hope this does not go unanswered beyond the close of business @ Stanford(PST).

[jadeshi]

Hey all,
Apologies for the inconvenience this has caused everyone. I have been checking over/monitoring the WS and project in question, and as far as I can tell 9201 is configured correctly to accept Maxwell GPUs work requests. I have also confirmed that this is not an issue of shortage of WU's, as there have been on the order of a hundred WU's consistently in the queue over the course of the day. I haven't touched this project and its settings for the several months it has been on full FAH and my understanding is that this issue of Maxwell incompatibility with 9201 only arose recently (i.e. today, someone correct me if this is not the case!). Thus, my best guess at this point is that it could be an AS issue, but unfortunately I have very little knowledge of and control over that.

We're currently discussing internally about it and we'll give you guys an update when this is resolved.

[rwh202]

Anyone that commented about getting low points on core 18 should have considered that getting no work units was worth even less points.

Windows still gets core_15. As far as I know there never was core_18 for Maxwell on Linux.

Just for what it's worth, I've got 4 9201s on Kepler but none on maxwell when I checked this morning (running 8 of each at the moment). Once the assignment is sorted maybe they should be further prioritised away from other cards to give maxwell and especially Linux enough work.