I am working with Intel QuickSync based GPU accelerated transcoding software (created by various sources), and I need to find a way to monitor the following for the Intel GPU on each server:
- Total GPU load
- Maximum number of potential GPU QuickSync threads supported for the hardware (CPU/GPU version)
- Number of current GPU QuickSync threads active
I want to do this so that the transcoding workload can be "load balanced" across a pool of servers, and to know when the GPU for a selected server is at it's maximum load.
I was told that the "Intel Metrics Framework" (within the "Intel Platform Analysis Library") could be used to do this, but I can not find any documentation on this specific data items that I would have to query for.
Anyone have any example on how to do this, or know of specific documentation that can help me???