[rsf-user] Compile CUDA program on Slurm server

0 views
Skip to first unread message

RONCORONI GIACOMO

unread,
Feb 13, 2022, 10:23:05 AM2/13/22
to rsf-...@lists.sourceforge.net
Hello everybody,

Please apologies if this question has been asked elsewhere - I did not find it anywhere-.

I'm trying to use sfgenshots on a server machine with GPUs, but I'm not able to compile it.

In the Slurm server I do not have sudo privileges, but I solved it by installing everything with anaconda.
Everything works except for gpu optimized programs.

Cuda is already installed in the machine and I have loaded before compiling on a compute node with 4 GPUs.
The problem is that when I run:

./configure --prefix=/m100/home/userexternal/groncor1/madagascar/src/

It turns out:
checking for CUDA ... (cached) no

I have tried to modify manully "CUDA_TOOLKIT_PATH = None" in config.py, setting it to: "/m100/prod/opt/compilers/cuda/11.0/none/bin/"

but i still get the same error after
make; make install

The error I get is:

(madagascar) [groncor1@login01 ~]$ sfgenshots

/m100/home/userexternal/groncor1/madagascar/src/bin/sfgenshots is not installed.

                             Please check if CUDA is not missing and is installed correctly.
                             Rerun Madagascar installation after it has been fixed.
So it seams it did not find nvcc or some other cuda-related compilers.

How could I solve this issue?

Few server info:
*********************************************************************************
* Welcome to MARCONI100 Cluster /                                              
*            IBM Power AC922 (Whiterspoon) -                                   
*            Red Hat Enterprise Linux Server release 8.1 (Ootpa)               
*                                                                              
* 980 compute nodes with:                                                     
*       - 2x16 cores IBM POWER9 AC922 at 3.1 GHz              
*       - 4 x NVIDIA Volta V100 GPUs, Nvlink 2.0, 16GB         
*       - 256 GB RAM                                                           

*********************************************************************************

Thank you very much and sorry if it had been asked elsewhere,
Sincerely,
Giacomo Roncoroni

_____________________________________________________________

Giacomo Roncoroni

Ph.D. student

Department of Mathematics and Geosciences

University of Trieste

Via Weiss 1, 34128, Trieste (IT)


Reply all
Reply to author
Forward
0 new messages