New Assignment Server feedback/problem

[Kjetil]

Thats is the same for me on 750Ti 3x and 980 3x. On my 780Ti x5 is working fine on P9406.

[kimben777]

I have four 780 ti's (kepler of course) and have been folding on 9406 wu's since yesterday with no problems, before the new AS I was folding 13000 and 13001 wu's for months also with no problems, I have a hard time believing it's bad wu's.

[Breach]

I have four 780 ti's (kepler of course) and have been folding on 9406 wu's since yesterday with no problems, before the new AS I was folding 13000 and 13001 wu's for months also with no problems, I have a hard time believing it's bad wu's.

These only cause problems on Maxwells, so perhaps the projects are not "faulty" as such.

[RipD]

problematic projects: 9406, 13000 and 13001 which all fail (at least on Maxwell, perhaps they fare better on Keplers).

Same for me: have gotten 9406, 13000, and 13001 for past day or so and all have failed. Will fire it back up to see if I get something different.

[Biffa]

Do most people feel like it's an issue w/specific projects (and that the new AS is leading to more assignments of them)? That would make sense. We can bring the faulty projects back to beta if there is enough evidence that they are faulty. We just shouldn't bring too many projects back to beta if that's not needed since otherwise there won't be any WUs of course, so we'll have to tread carefully here.

Definitely a Maxwell only issue as far as I can ascertain.

People with 750's and 970/980's with no changes to their hardware are suddenly getting problems both with -advanced and -beta projects using Core17/18

Core_18 Projects 10473, 10472, 10471 I didn't get a 10470 because I didn't want to sit generating errors.

Core_17 Projects 9406 is the only one I have seen fail on my Maxwells

[Kjetil]

Thats correct, my 5 780Ti is running fine. My 6 maxwell is not.

[widsss]

It's only maxwell, but something changed on Stanford's side. My cards are the same, my driver is the same, Maxwell did sucessfully fold many WUs before saturday. Surely Pande can pinpoint the time frame and what changes were made prior to the bad work units(?).

[tbk-aracthebold]

Hi,

I've been folding the 9406 just fine on my gtx 660ti and gtx 760. I'm still using the 327 drivers. Here is a part of both slots log.

Code: Select all

21:08:27:WU00:FS02:0x17:*********************** Log Started 2014-10-05T21:08:27Z ***********************
21:08:27:WU00:FS02:0x17:Project: 9406 (Run 209, Clone 1, Gen 25)
21:08:27:WU00:FS02:0x17:Unit: 0x0000002b0a3b1e5c533def7dd828e1ed
21:08:27:WU00:FS02:0x17:CPU: 0x00000000000000000000000000000000
21:08:27:WU00:FS02:0x17:Machine: 2
21:08:27:WU00:FS02:0x17:Digital signatures verified
21:08:27:WU00:FS02:0x17:Folding@home GPU core17
21:08:27:WU00:FS02:0x17:Version 0.0.55
21:08:28:WU00:FS02:0x17:  Found a checkpoint file
21:12:51:WU00:FS02:0x17:Completed 1050000 out of 2000000 steps (52%)
21:12:51:WU00:FS02:0x17:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
21:16:23:WU00:FS02:0x17:Completed 1060000 out of 2000000 steps (53%)
21:22:31:WU00:FS02:0x17:Completed 1080000 out of 2000000 steps (54%)
21:28:39:WU00:FS02:0x17:Completed 1100000 out of 2000000 steps (55%)
21:35:14:WU00:FS02:0x17:Completed 1120000 out of 2000000 steps (56%)
21:41:22:WU00:FS02:0x17:Completed 1140000 out of 2000000 steps (57%)
21:47:58:WU00:FS02:0x17:Completed 1160000 out of 2000000 steps (58%)
21:54:06:WU00:FS02:0x17:Completed 1180000 out of 2000000 steps (59%)
22:00:16:WU00:FS02:0x17:Completed 1200000 out of 2000000 steps (60%)
22:06:52:WU00:FS02:0x17:Completed 1220000 out of 2000000 steps (61%)
22:13:01:WU00:FS02:0x17:Completed 1240000 out of 2000000 steps (62%)
22:19:37:WU00:FS02:0x17:Completed 1260000 out of 2000000 steps (63%)
22:25:46:WU00:FS02:0x17:Completed 1280000 out of 2000000 steps (64%)
22:31:55:WU00:FS02:0x17:Completed 1300000 out of 2000000 steps (65%)
22:38:32:WU00:FS02:0x17:Completed 1320000 out of 2000000 steps (66%)
22:44:41:WU00:FS02:0x17:Completed 1340000 out of 2000000 steps (67%)
22:51:16:WU00:FS02:0x17:Completed 1360000 out of 2000000 steps (68%)
22:57:26:WU00:FS02:0x17:Completed 1380000 out of 2000000 steps (69%)
23:03:34:WU00:FS02:0x17:Completed 1400000 out of 2000000 steps (70%)
23:10:10:WU00:FS02:0x17:Completed 1420000 out of 2000000 steps (71%)
23:16:18:WU00:FS02:0x17:Completed 1440000 out of 2000000 steps (72%)
23:22:55:WU00:FS02:0x17:Completed 1460000 out of 2000000 steps (73%)
23:29:04:WU00:FS02:0x17:Completed 1480000 out of 2000000 steps (74%)
23:35:13:WU00:FS02:0x17:Completed 1500000 out of 2000000 steps (75%)
23:41:47:WU00:FS02:0x17:Completed 1520000 out of 2000000 steps (76%)
23:47:55:WU00:FS02:0x17:Completed 1540000 out of 2000000 steps (77%)
23:54:27:WU00:FS02:0x17:Completed 1560000 out of 2000000 steps (78%)
00:00:36:WU00:FS02:0x17:Completed 1580000 out of 2000000 steps (79%)
00:06:44:WU00:FS02:0x17:Completed 1600000 out of 2000000 steps (80%)
00:13:16:WU00:FS02:0x17:Completed 1620000 out of 2000000 steps (81%)
00:19:24:WU00:FS02:0x17:Completed 1640000 out of 2000000 steps (82%)

Code: Select all

23:35:22:WU04:FS03:0x17:*********************** Log Started 2014-10-05T23:35:22Z ***********************
23:35:22:WU04:FS03:0x17:Project: 9406 (Run 287, Clone 0, Gen 92)
23:35:22:WU04:FS03:0x17:Unit: 0x0000009f0a3b1e5c533dfdd50826da28
23:35:22:WU04:FS03:0x17:CPU: 0x00000000000000000000000000000000
23:35:22:WU04:FS03:0x17:Machine: 3
23:35:22:WU04:FS03:0x17:Digital signatures verified
23:35:22:WU04:FS03:0x17:Folding@home GPU core17
23:35:22:WU04:FS03:0x17:Version 0.0.55
23:35:22:WU04:FS03:0x17:  Found a checkpoint file
23:38:46:WU04:FS03:0x17:Completed 1200000 out of 2000000 steps (60%)
23:38:46:WU04:FS03:0x17:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
23:45:09:WU04:FS03:0x17:Completed 1220000 out of 2000000 steps (61%)
23:51:10:WU04:FS03:0x17:Completed 1240000 out of 2000000 steps (62%)
23:57:35:WU04:FS03:0x17:Completed 1260000 out of 2000000 steps (63%)
00:03:38:WU04:FS03:0x17:Completed 1280000 out of 2000000 steps (64%)
00:09:39:WU04:FS03:0x17:Completed 1300000 out of 2000000 steps (65%)
00:16:02:WU04:FS03:0x17:Completed 1320000 out of 2000000 steps (66%)
00:22:04:WU04:FS03:0x17:Completed 1340000 out of 2000000 steps (67%)

[Kjetil]

Nothing is wrong whit your log as i can see, so more info please.

[tbk-aracthebold]

Nothing is wrong whit your log as i can see, so more info please.

What info do you want? I was just showing that my machine is not having issues folding the 9406 units. It also folded a few of the core 18 units with no issue. As I stated I am still using the 327.23 drivers, If you are having failures you might want to see if going back to the 327.23 drivers fix the failures.

wonder if this has anything to do with my being able to fold the 9406's

GPU 0: NVIDIA:3 GK104 [GeForce GTX 660 Ti]
GPU 1: NVIDIA:3 GK104 [GeForce GTX 760]

all the logs show the gpu's as gen 4 gk107's

[Kjetil]

Nothing is wrong whit your log as i can see, so more info please.

What info do you want? I was just showing that my machine is not having issues folding the 9406 units. It also folded a few of the core 18 units with no issue. As I stated I am still using the 327.23 drivers, If you are having failures you might want to see if going back to the 327.23 drivers fix the failures.

Sorry, my bad. but you are the first one so post you did not have a problems. So way post if you do not have any problems?
You do not running maxwell, i have 6 maxwell all bad, my 5 780Ti is not. Sorry.

[snapshot]

I have three different systems with GTX 750Ti (Maxwell) video cards, the 7.4.4 client, Windows 7, and the Nvidia 340.52 drivers.

Upgrade to the latest drivers

(I've added that to the announcement.)

1. It doesn't make any difference.
2. They give a 10% drop in PPD so aren't fit for purpose. That may be better than no PPD but see point 1......

[Kjetil]

My maxwell is running now.

[VijayPande]

ok, we'll set 9406, 13000 and 13001 to be beta only for Maxwell. That should keep them away from the Maxwell GPUs and still give us beta team feedback for what's going wrong w/those WUs.

[Kjetil]

Okay, running 78xx and 9202 now. Thanks.
Edit: Can you fix hfm.net to? It show 0 points and core unknown on 78xx.