Re: [slurm-users] nvml autodetect is ignoring gpus

2021-12-01 Thread Benjamin Nacar
Confirmed that adding just the "Gres=" bit in slurm.conf works. That's what I get for reading the documentation too fast... thanks all! ~~ bnacar On Wed, 1 Dec 2021 14:05:09 +0100 Quirin Lohr wrote: > Hi, > > you still need to specify the gpus in the node definition in slurm.conf. > At least

Re: [slurm-users] nvml autodetect is ignoring gpus

2021-12-01 Thread Quirin Lohr
Hi, you still need to specify the gpus in the node definition in slurm.conf. At least the number, perhaps even the type reported by nvml must match the node definition. (Gres=gpu:geforce_gtx_1080:4) I think the error message can be ignored, the 1080 just does not support this feature. Am

Re: [slurm-users] nvml autodetect is ignoring gpus

2021-12-01 Thread Fernando Guillén Camba
I also compiled Slurm 20.11.8 to have GPU support in AlmaLinux 8.4 but don't have any problem with  NVML detecting our A100s. ¿Maybe the NVML library version used for Slurm compilation has to match the library version of the compute node where the GPU is? Also, I see that you're using Geforce

Re: [slurm-users] nvml autodetect is ignoring gpus

2021-11-30 Thread Diego Zuccato
Il 30/11/2021 16:12, Benjamin Nacar ha scritto: However, the version of Slurm in the standard debian repositories was apparently not compiled on a system with the necessary Nvidia library installed, That's not a good news :( I have a GPU node arriving by the end of the year. Does it only impact

[slurm-users] nvml autodetect is ignoring gpus

2021-11-30 Thread Benjamin Nacar
Hi, We're trying to use Slurm's built-in Nvidia GPU detection mechanism to avoid having to specify GPUs explicitly in slurm.conf and gres.conf. We're running Debian 11, and the version of Slurm available for Debian 11 is 20.11. However, the version of Slurm in the standard debian repositories w