Clara Deploy SDK is being consolidated into Clara Holoscan SDK

More info https://catalog.ngc.nvidia.com/orgs/nvidia/teams/clara/collections/claradeploy

Clara Hardware Qualification Suite

This asset requires the Clara Deploy SDK. Follow the instructions on the Clara Ansible page to install the Clara Deploy SDK.

Overview

The Clara Hardware Qualification Suite is a benchmarking and scoring utility that can be used to validate that hardware meets the minimum requirements for Clara. It can also help diagnose run-time performance issues.

Set of metrics calculated

Information Tests

NVIDIA Driver details
GPU information
CPU information
System-memory capacity and availability
Persistent-storage capacity and availability

Exercise Tests

System-memory read/write performance
CPU computational performance (single thread)
CPU computational performance (multi-threaded)
Persistent storage read/write performance
GPU compute performance

Running the Qualification Suite Inside a Container

sudo docker run -it --rm -v `pwd`:/output --name clara-hqs clara-hqs[:version]

# or, you can add `--entrypoint "./run -i -m -s -g -c single multi -e"` in the above command after `--rm`

# usage: ./run [-h] [-i] [-e] [-m] [-c args [args ...]] [-s] [-g]
# [-d ]

# optional arguments:
# -h, --help show this help message and exit
# -i, --info get hardware information
# -e, --export export output to a json file
# -m, --memory get memory read write performance
# -c args [args ...], --cpu args [args ...]
# get cpu performance by specifying space separated list
# of types of threading [args: single, multi]
# -s, --storage get Persistent storage read write performance
# -g, --gpu get gpu performance
# -d , --dir 
# cuda toolkit samples directory

License

An End User License Agreement is included with the product. By pulling and using the Clara Deploy asset on NGC, you accept the terms and conditions of these licenses.

Clara Deploy Hardware Qualification Suite Pipeline [Deprecated]