a kit to run on AHP to consume various resources
CPU: Number of CPU cores to stress (e.g., 4 or 0 for all cores). RAM_WORKERS: Number of memory allocation threads (e.g., 2). RAM_BYTES: Size of RAM per worker to consume (e.g., 2G or 80%). GPU_TIME: Duration in seconds to run the GPU matrix calculation stress test. TIME: Duration for the CPU/RAM stress test (e.g., 60s, 2m).
Run with environment variables:You must pass the --gpus all flag so Docker can access your host GPU.
docker run --rm --gpus all
-e CPU="4"
-e RAM_WORKERS="2"
-e RAM_BYTES="2G"
-e GPU_TIME="60"
-e TIME="60s"
hardware-stress-test