One nV GPU folding and one not

[dmoimas_geo]

Hello everybody,

this is my first post.

I recently updated FAHClient on my linux box (HP Proliant DL580 G5) which is running 3 slots: CPU (:12), 2 GPUs. Linux Ubuntu 18.04.4, hwe kernel, nVidia driver 390.132.
One GPU is a Tesla C2075 (GF110) and the other is a Tesla C2050 (GF100). Both supports only OpenCL 1.1 (due to nV OpenCl support in 390.x driver release for Fermi) but are DP capable.

Before upgrading to 7.6.13 I had no issues folding with both GPUs.
Now I only get C2075 to fold, while C2050 is stuck at "assigned to work server 192.0.2.1".

"13:40:57:WU04:FS03:Assigned to work server 192.0.2.1
13:40:57:WU04:FS03:Requesting new work unit for slot 03: READY gpu:0:GF100 [Tesla C2050 / C2070] from 192.0.2.1
13:40:57:WU04:FS03:Connecting to 192.0.2.1:8080"

This is my GPUs.txt for my GPUs... seems the same.
fah@foldingathome:~$ cat /var/lib/fahclient/GPUs.txt | grep "Tesla C20"
0x10de:0x06d1:2:2:GF100 [Tesla C2050 / C2070]
0x10de:0x108e:2:2:GF110GL [Tesla C2090]
0x10de:0x1096:2:2:GF110 [Tesla C2075 Dual-Slot CPM]

I tried deleting GPU slot and recreating (was 1, now it's 3) but no success.

Thanks in advance.

Best regards,
Diego

[Joe_H]

See this post - viewtopic.php?f=74&t=35220&p=334019#p334019

Folding on the Tesla C2050 should get reenabled in a bit.

[dmoimas_geo]

Hello Joe,

thanks. To be sure, I sent bruce a PM.

Best regards,
Diego

[bruce]

That GPU is supported as species 2.

Post the log showing the difficulties you're having.

[dmoimas_geo]

Hello Bruce,

problem got fixed automatically. Now it's folding without issues.

Before I had that "assign1" server was always assigning slot (C2050) to 192.0.2.1

BR,
Diego