Docker image for the Multi-GPU CUDA stress test from http://wili.cc/blog/gpu-burn.html
Below commands assume that you have the appropriate version of the NVIDIA Container Toolkit installed.
Default test duration is 60 seconds. If needed, change to a different value.
docker run --gpus all --rm oguzpastirmaci/gpu-burn <test duration in seconds>
nvidia-docker run --rm oguzpastirmaci/gpu-burn <test duration in seconds>