Discussion:
[OMPI users] openmpi-master-201708110239-03544d7: NVIDIA: no NVIDIA devices found
Siegmar Gross
2017-08-14 07:46:50 UTC
Permalink
Hi,

I have installed openmpi-master-201708110239-03544d7 and openmpi-2.1.2rc1
on my "SUSE Linux Enterprise Server 12.2 (x86_64)" with Sun C 5.15 and
gcc-5.3.0. "mpiexec" from openmpi-master reports "NVIDIA: no NVIDIA
devices found" if a machine isn't equipped with a Nvidia device.


loki fd1026 105 mpiexec --host nfs1 hostname
NVIDIA: no NVIDIA devices found
nfs1
loki fd1026 106 which mpiexec
/usr/local/openmpi-master_64_gcc/bin/mpiexec


loki fd1026 110 mpiexec --host nfs1 hostname
nfs1
loki fd1026 111 which mpiexec
/usr/local/openmpi-2.1.2_64_gcc/bin/mpiexec


Both installations support CUDA.

loki fd1026 112 find /usr/local/openmpi-master_64_gcc/lib64 -name '*cuda*'
/usr/local/openmpi-master_64_gcc/lib64/libmca_common_cuda.so.0.0.0
/usr/local/openmpi-master_64_gcc/lib64/libmca_common_cuda.so.0
/usr/local/openmpi-master_64_gcc/lib64/openmpi/mca_coll_cuda.la
/usr/local/openmpi-master_64_gcc/lib64/openmpi/mca_btl_smcuda.so
/usr/local/openmpi-master_64_gcc/lib64/openmpi/mca_btl_smcuda.la
/usr/local/openmpi-master_64_gcc/lib64/openmpi/mca_coll_cuda.so
/usr/local/openmpi-master_64_gcc/lib64/libmca_common_cuda.la
/usr/local/openmpi-master_64_gcc/lib64/libmca_common_cuda.so

loki fd1026 113 find /usr/local/openmpi-2.1.2_64_gcc/lib64 -name '*cuda*'
/usr/local/openmpi-2.1.2_64_gcc/lib64/libmca_common_cuda.so.20.10.0
/usr/local/openmpi-2.1.2_64_gcc/lib64/openmpi/mca_coll_cuda.la
/usr/local/openmpi-2.1.2_64_gcc/lib64/openmpi/mca_btl_smcuda.so
/usr/local/openmpi-2.1.2_64_gcc/lib64/openmpi/mca_btl_smcuda.la
/usr/local/openmpi-2.1.2_64_gcc/lib64/openmpi/mca_coll_cuda.so
/usr/local/openmpi-2.1.2_64_gcc/lib64/libmca_common_cuda.la
/usr/local/openmpi-2.1.2_64_gcc/lib64/libmca_common_cuda.so.20
/usr/local/openmpi-2.1.2_64_gcc/lib64/libmca_common_cuda.so


I would be grateful, if somebody can fix the problem. Thank you very
much for any help in advance.


Kind regards

Siegmar
Sylvain Jeaugey
2017-08-14 16:55:19 UTC
Permalink
Hi Siegmar,

This has been fixed in the driver some time ago. Getting the latest
driver should solve your problem.

You can check the driver version with nvidia-smi, then go to
http://www.nvidia.com/Download/index.aspx to get the latest.

Sylvain
Post by Siegmar Gross
Hi,
I have installed openmpi-master-201708110239-03544d7 and openmpi-2.1.2rc1
on my "SUSE Linux Enterprise Server 12.2 (x86_64)" with Sun C 5.15 and
gcc-5.3.0. "mpiexec" from openmpi-master reports "NVIDIA: no NVIDIA
devices found" if a machine isn't equipped with a Nvidia device.
loki fd1026 105 mpiexec --host nfs1 hostname
NVIDIA: no NVIDIA devices found
nfs1
loki fd1026 106 which mpiexec
/usr/local/openmpi-master_64_gcc/bin/mpiexec
loki fd1026 110 mpiexec --host nfs1 hostname
nfs1
loki fd1026 111 which mpiexec
/usr/local/openmpi-2.1.2_64_gcc/bin/mpiexec
Both installations support CUDA.
loki fd1026 112 find /usr/local/openmpi-master_64_gcc/lib64 -name '*cuda*'
/usr/local/openmpi-master_64_gcc/lib64/libmca_common_cuda.so.0.0.0
/usr/local/openmpi-master_64_gcc/lib64/libmca_common_cuda.so.0
/usr/local/openmpi-master_64_gcc/lib64/openmpi/mca_coll_cuda.la
/usr/local/openmpi-master_64_gcc/lib64/openmpi/mca_btl_smcuda.so
/usr/local/openmpi-master_64_gcc/lib64/openmpi/mca_btl_smcuda.la
/usr/local/openmpi-master_64_gcc/lib64/openmpi/mca_coll_cuda.so
/usr/local/openmpi-master_64_gcc/lib64/libmca_common_cuda.la
/usr/local/openmpi-master_64_gcc/lib64/libmca_common_cuda.so
loki fd1026 113 find /usr/local/openmpi-2.1.2_64_gcc/lib64 -name '*cuda*'
/usr/local/openmpi-2.1.2_64_gcc/lib64/libmca_common_cuda.so.20.10.0
/usr/local/openmpi-2.1.2_64_gcc/lib64/openmpi/mca_coll_cuda.la
/usr/local/openmpi-2.1.2_64_gcc/lib64/openmpi/mca_btl_smcuda.so
/usr/local/openmpi-2.1.2_64_gcc/lib64/openmpi/mca_btl_smcuda.la
/usr/local/openmpi-2.1.2_64_gcc/lib64/openmpi/mca_coll_cuda.so
/usr/local/openmpi-2.1.2_64_gcc/lib64/libmca_common_cuda.la
/usr/local/openmpi-2.1.2_64_gcc/lib64/libmca_common_cuda.so.20
/usr/local/openmpi-2.1.2_64_gcc/lib64/libmca_common_cuda.so
I would be grateful, if somebody can fix the problem. Thank you very
much for any help in advance.
Kind regards
Siegmar
_______________________________________________
users mailing list
https://lists.open-mpi.org/mailman/listinfo/users
-----------------------------------------------------------------------------------
This email message is for the sole use of the intended recipient(s) and may contain
confidential information. Any unauthorized review, use, disclosure or distribution
is prohibited. If you are not the intended recipient, please contact the sender by
reply email and destroy all copies of the original message.
-----------------------------------------------------------------------------------
Loading...