gpu-burn

Multi-GPU CUDA stress test http://wili.cc/blog/gpu-burn.html

Building

To build GPU Burn:

make

To remove artifacts built by GPU Burn:

make clean

GPU Burn builds with a default Compute Capability of 5.0. To override this with a different value:

make COMPUTE=<compute capability value>

CFLAGS can be added when invoking make to add to the default list of compiler flags:

make CFLAGS=-Wall

LDFLAGS can be added when invoking make to add to the default list of linker flags:

make LDFLAGS=-lmylib

NVCCFLAGS can be added when invoking make to add to the default list of nvcc flags:

make NVCCFLAGS=-ccbin <path to host compiler>

CUDAPATH can be added to point to a non standard install or specific version of the cuda toolkit (default is /usr/local/cuda):

make CUDAPATH=/usr/local/cuda-<version>

CCPATH can be specified to point to a specific gcc (default is /usr/bin):

make CCPATH=/usr/local/bin

Usage

GPU Burn
Usage: gpu_burn [OPTIONS] [TIME]

-m 90 Gpu percent
-d	Use doubles
-tc	Use Tensor cores
-h	Show this help message

Example:
gpu-burn -d 3600
gpu_burn -m 96% -d -tc 60

Using Docker

docker build -t gpu-burn .

docker run -it --rm --gpus all gpu-burn