BAD_WORK_UNIT et GPU AMD

[JWhy]

Pas mal BAD_WORK_UNIT de mon coté ces dernieres heures/derniers jours (certaines également FAULTY chez d'autres plieurs, cf. https://apps.foldingathome.org/wu )

Code :

01:53:33:WU00:FS00:Sending unit results: id:00 state:SEND error:NO_ERROR project:11777 run:0 clone:23631 gen:17 core:0x22 unit:0x0000001c287234c95e7749fffffe0e1c
09:10:52:WU01:FS00:Sending unit results: id:01 state:SEND error:NO_ERROR project:11779 run:0 clone:1948 gen:31 core:0x22 unit:0x0000002d0d5a98395e73c5c57518e51b
16:09:58:WU00:FS00:Sending unit results: id:00 state:SEND error:NO_ERROR project:11779 run:0 clone:437 gen:44 core:0x22 unit:0x000000360d5a98395e73c5de6d83a3fa
17:21:57:WU01:FS00:Sending unit results: id:01 state:SEND error:FAULTY project:11776 run:0 clone:4209 gen:9 core:0x22 unit:0x0000001d287234c95e73c45dd1e3c4fa
18:10:23:WU00:FS00:Sending unit results: id:00 state:SEND error:FAULTY project:11764 run:0 clone:2502 gen:40 core:0x22 unit:0x0000003e80fccb0a5e6d8271bb659f30
00:53:47:WU01:FS00:Sending unit results: id:01 state:SEND error:FAULTY project:11781 run:0 clone:4816 gen:29 core:0x22 unit:0x000000310d5a98395e73c4f893e38b3a
01:02:49:WU00:FS00:Sending unit results: id:00 state:SEND error:FAULTY project:11764 run:0 clone:3187 gen:33 core:0x22 unit:0x0000003780fccb0a5e6d85da25ea0604
09:02:17:WU01:FS00:Sending unit results: id:01 state:SEND error:NO_ERROR project:11778 run:0 clone:29892 gen:23 core:0x22 unit:0x00000025287234c95e77482e18da6e1d
09:02:51:WU00:FS00:Sending unit results: id:00 state:SEND error:FAULTY project:11759 run:0 clone:2960 gen:37 core:0x22 unit:0x0000003f80fccb0a5e6d7ca7854675bc
09:06:47:WU03:FS00:Sending unit results: id:03 state:SEND error:FAULTY project:11776 run:0 clone:12249 gen:19 core:0x22 unit:0x00000026287234c95e74333c7fa22e47
01:53:33:WU00:FS00:Sending unit results: id:00 state:SEND error:NO_ERROR project:11777 run:0 clone:23631 gen:17 core:0x22 unit:0x0000001c287234c95e7749fffffe0e1c
09:10:52:WU01:FS00:Sending unit results: id:01 state:SEND error:NO_ERROR project:11779 run:0 clone:1948 gen:31 core:0x22 unit:0x0000002d0d5a98395e73c5c57518e51b
16:09:58:WU00:FS00:Sending unit results: id:00 state:SEND error:NO_ERROR project:11779 run:0 clone:437 gen:44 core:0x22 unit:0x000000360d5a98395e73c5de6d83a3fa
17:21:57:WU01:FS00:Sending unit results: id:01 state:SEND error:FAULTY project:11776 run:0 clone:4209 gen:9 core:0x22 unit:0x0000001d287234c95e73c45dd1e3c4fa
18:10:23:WU00:FS00:Sending unit results: id:00 state:SEND error:FAULTY project:11764 run:0 clone:2502 gen:40 core:0x22 unit:0x0000003e80fccb0a5e6d8271bb659f30
00:53:47:WU01:FS00:Sending unit results: id:01 state:SEND error:FAULTY project:11781 run:0 clone:4816 gen:29 core:0x22 unit:0x000000310d5a98395e73c4f893e38b3a
01:02:49:WU00:FS00:Sending unit results: id:00 state:SEND error:FAULTY project:11764 run:0 clone:3187 gen:33 core:0x22 unit:0x0000003780fccb0a5e6d85da25ea0604
10:24:49:WU00:FS00:Sending unit results: id:00 state:SEND error:FAULTY project:11776 run:0 clone:8419 gen:28 core:0x22 unit:0x0000002f287234c95e743375b03a9d44
10:29:49:WU01:FS00:Sending unit results: id:01 state:SEND error:FAULTY project:11781 run:0 clone:5226 gen:26 core:0x22 unit:0x000000320d5a98395e758911e444f866
10:37:24:WU00:FS00:Sending unit results: id:00 state:SEND error:FAULTY project:11781 run:0 clone:6558 gen:41 core:0x22 unit:0x000000420d5a98395e7588fe03d35e85

Avec toujours le même msg d'erreur.

Code :

01:02:48:WU00:FS00:0x22:ERROR:exception: Error invoking kernel sortShortList: clEnqueueNDRangeKernel (-5)
01:02:48:WU00:FS00:0x22:Folding@home Core Shutdown: BAD_WORK_UNIT
01:02:49:WARNING:WU00:FS00:FahCore returned: BAD_WORK_UNIT (114 = 0x72)

Je ne suis a priori pas le seul :
- https://foldingforum.org/viewtopic.php?f=19&t=32301
- https://foldingforum.org/viewtopic.php?f=74&t=32991
- https://foldingforum.org/viewtopic.php?f=81&t=32771

Ça semble lié à l'implémentation d'OpenCL sur certains drivers / certaines cartes AMD (j'ai une RX 460)

Code :

09:02:51:WU03:FS00:Requesting new work unit for slot 00: READY gpu:0:Baffin XT [Radeon RX 460] from 40.114.52.201

Someone else looked at this kind of error, it appears from their research that there is a limit set in the AMD OpenCL support that may be causing problems if the size of the WU exceeds some large number of atoms.  This will need to be checked into, it is not clear whether this just applies to some version of the drivers, is actually the problem, and so on.  The projects running into this error may get restricted from assignments to AMD GPU's at some point, a report was passed on to the F@h group.  Or the fix might be something else, too soon to be more definite.

Researchers are looking into disabling those projects on AMD GPUs until fix has been found.
Thank you for understanding

Je regarderai si une MàJ des drivers existe dans la journée, vu que je continue à avoir ce type de WU.