Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Determine and display various versions of interest #521

Open
wants to merge 4 commits into
base: main
Choose a base branch
from

Conversation

iankouls-aws
Copy link
Contributor

Adds bash script which determines and prints various versions of interest.
Inventory of installed components and versions is useful to determine whether prerequisites are satisfied in various scenarios.
This PR does not depend on any other PRs or fix a reported issue.

@iankouls-aws
Copy link
Contributor Author

Sample output:

[root@ip-172-16-76-105 tools]# ./versions.sh

OS info:
NAME="Amazon Linux"
VERSION="2"
ID="amzn"
ID_LIKE="centos rhel fedora"

Linux family:
Linux

Linux Kernel version:
5.10.228-219.884.amzn2.x86_64

nvidia-smi:
/usr/bin/nvidia-smi

NVIDIA versions:
DRIVER version      : 550.127.05
CUDA Version        : 12.4
NCCL version: 2.22.3
AWS OFI NCCL version: 1.13.0
NVIDIA GDS:
 GDS release version: 1.7.2.10
 libcufile version: 2.12
 Platform: x86_64

Lustre client version:
/usr/bin/yum
2.12.8-4.amzn2

EFA Installer version:
efa-2.13.0-1.amzn2.x86_64

Libfabric version:
1.22.0amzn4.0

if [ "$?" == "0" ]; then
echo ""
echo "NVIDIA versions:"
nvidia-smi --version | grep DRIVER
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
nvidia-smi --version | grep DRIVER
nvidia-smi --query-gpu=driver_version --format=csv,noheader --id=0

echo ""
echo "NVIDIA versions:"
nvidia-smi --version | grep DRIVER
nvidia-smi --version | grep CUDA
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
nvidia-smi --version | grep CUDA
nvcc --version | sed -n 's/^.*release \([0-9]\+\.[0-9]\+\).*$/\1/p

nvidia smi provices the cuda version support by the nvidia driver not the actual version of cuda

nvidia-smi --version | grep DRIVER
nvidia-smi --version | grep CUDA

NCCL_LIB=$(ls /usr/local/cuda/lib/libnccl.so.*.*.*)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It points toward the default path of cuda install. However there can be many version cuda installed.

LUSTRE_CLIENT_VERSION="not found"
which yum
if [ "$?" == "0" ]; then
LUSTRE_CLIENT_VERSION=$(yum list lustre-client | grep lustre-client | awk '{print $2}')
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You are checking installed version, not the actual version loaded.

Suggested change
LUSTRE_CLIENT_VERSION=$(yum list lustre-client | grep lustre-client | awk '{print $2}')
modinfo lustre | grep 'version:' | head -n 1 | awk '{print $2}'

Copy link
Contributor

@mhuguesaws mhuguesaws left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Left comments.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants