Skip to content
This repository has been archived by the owner on Jul 22, 2024. It is now read-only.

Creating PowerAI Verification Tool. #10

Open
wants to merge 2 commits into
base: master
Choose a base branch
from

Conversation

pvemuri1
Copy link

@pvemuri1 pvemuri1 commented Dec 1, 2018

This tool validates the below:

  • Is pre-reqs installed correcly
  • Is License accepted
  • Is PowerAI installed
  • Is frameworks Operational

This tool validates the below:
* Is pre-reqs installed correcly
* Is License accepted
* Is PowerAI installed
* Is frameworks Operational
required_version=$nvidia_version
installed_version=`nvidia-smi | grep NVIDIA-SMI | awk '{print $3}'`

function version_ge() { test "$(echo "$@" | tr " " "\n" | sort -rV | head -n 1)" == "$1"; }

Choose a reason for hiding this comment

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

This is checking if they are equal, but error message is saying that it is checking if they are greater than or equal. We want to check if it is equal, so the error message should be changed.

echo " RHEL7 powerle optional repos are enabled." 2>&1 >> "${LOGFILE}"
IS_OPTIONAL_REPO_ENABLED="YES"
else
echo " Optional repos are to be enabled to install PowerAI." 2>&1 | tee -a "${LOGFILE}"

Choose a reason for hiding this comment

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

Message should be:
" Optional repos must be enabled in order to install PowerAI."

echo " RHEL7 powerle extras repo is enabled." 2>&1 >> "${LOGFILE}"
IS_EXTRAS_REPO_ENABLED="YES"
else
echo " Extras repo is to be enabled to install PowerAI." 2>&1 | tee -a "${LOGFILE}"

Choose a reason for hiding this comment

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

Message should be:
" Extras repo must be enabled in order to install PowerAI."

echo " EPEL repo is enabled." 2>&1 >> "${LOGFILE}"
IS_EPEL_REPO_ENABLED="YES"
else
echo " EPEL repo is to be enabled to install PowerAI." 2>&1 | tee -a "${LOGFILE}"

Choose a reason for hiding this comment

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

Message should be:
" EPEL repo must be enabled in order to install PowerAI."

then
if version_ge $current_cuda_version $cuda_version;
then
echo " CUDA $current_cuda_version is greater than or equal to $cuda_version." 2>&1 >> "${LOGFILE}"

Choose a reason for hiding this comment

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

We are only checking for equality.

$LICENSE_BIN/check-powerai-license.sh
if [ $? -ne 0 ]
then
echo " PowerAI license is to be accpeted to activate the frameworks." 2>&1 | tee -a "${LOGFILE}"

Choose a reason for hiding this comment

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

Message should be:
"PowerAI license has not been accepted."

POWERAI_LICENSE_ACCEPTED="NO"
exit 1
else
echo " PowerAI license is accpeted." 2>&1 >> "${LOGFILE}"

Choose a reason for hiding this comment

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

Message should be:
PowerAI license is accepted.


validate_ddl(){
# Initialize report vars
#IS_DDL_INSTALLED="0"

Choose a reason for hiding this comment

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

Remove this line.

then
echo " DDL activated sucessfully" 2>&1 >> "${LOGFILE}"
IS_DDL_INSTALLED_REP="YES"
else

Choose a reason for hiding this comment

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

Fix indentation.

echo " DDL activated sucessfully" 2>&1 >> "${LOGFILE}"
IS_DDL_INSTALLED_REP="YES"
else
/opt/DL/tensorflow/bin/install_dependencies -y

Choose a reason for hiding this comment

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

We shouldn't be installing dependencies here.

Copy link

Choose a reason for hiding this comment

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

@pvemuri1 I suggest you to create our own conda environment pai_verify_env within this tool after validate_anaconda() is called. Then any framework verification/validation can actually perform install_dependencies with in this conda environment pai_verify_env and we can clean-up/remove this conda environment pai_verify_env as the end of execution of this verification tool.

@ravigumm
Copy link

ravigumm commented Dec 7, 2018

@pvemuri1 Please add copyright to these new files.

* Fixed indentation issues.
* Corrected messages.
* creating conda environment to install_dependencies in the conda environment.
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants