HGX2 Field Diagnostics Tool

Document Scope

This article covers how to utilize the HGX field diagnostics tool for troubleshooting HGX GPUs. These specialized GPUs are not easily removed from their installed chassis and as such cannot simply be placed into another system as a manner of A/B testing functionality. Instead, we rely on this tool to provide the insight necessary to determine the current state of the GPU.

Prerequisite

This tool needs to be on a bootable USB. If you already have the tool on the bootable USB, skip to 'Step 2'. The files can be found from this link:

https://exxact-support.s3-us-west-1.amazonaws.com/Test+Folder/hgx2-diagos_18.07.1-03_bundle.zip

Step 1: Create a bootable USB

I used Rufus 2.18, but basically any Rufus will work as long as you uncheck the "Create a bootable disk using..." options

Example of Rufus settings

Step 2: Boot to the USB

I have a PNY USB, and I used the UEFI option.

Step 3: Unpack Field Diagnostic Tool files

Use command below to unpack fie Field Diagnostics Tool files:

tar xfz 629-FKD03-2887-510.tgz

You should land into a command prompt: 

Step 4: Change directory into the unpacked files

Use command below to unpack fie Field Diagnostics Tool files:

cd xfz 629-FKD03-2887-510

Example of failed test:

Step 5: Provide logs generated by the test

Power off system once tests complete, remove the USB, and you can plug it in on your personal computer and find the logs under:

USB (Example) DGX2 (D:) → home → 629-FLD03-2887-510 → Logs → zipped files

The zipped files for the tests ran.