The testing was to ensure that our backend GPUs run at optimal temperatures. This will improve the concurrent requests being processed at a given time.