# GrabGPU **Repository Path**: tju_hfut_sym/GrabGPU ## Basic Information - **Project Name**: GrabGPU - **Description**: No description available - **Primary Language**: Unknown - **License**: Not specified - **Default Branch**: master - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 0 - **Created**: 2021-12-30 - **Last Updated**: 2021-12-30 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README # Grab the GPUs to run your own code! ## Download (downward compatibility) **CUDA 10.1:** ```shell wget https://github.com/godweiyang/GrabGPU/releases/download/v1.0.0/gg_cu101 ``` **CUDA 11.0:** ```shell wget https://github.com/godweiyang/GrabGPU/releases/download/v1.0.0/gg_cu110 ``` **CUDA 11.2:** ```shell wget https://github.com/godweiyang/GrabGPU/releases/download/v1.0.0/gg_cu112 ``` ## Compile the source code ```shell nvcc gg.cu -o gg ``` ## Run default script **Usage:** ```shell ./gg ``` **Example:** Occupy 16 GB GPU memory for 24 hours using GPU 0, 1, 2, 3 to run default script. ```shell ./gg 16 24 0,1,2,3 ``` ## Run your own script **Usage:** ```shell ./gg ``` **Example:** Occupy 16 GB GPU memory using GPU 0, 1, 2, 3 to run your own `run.sh`. Note that the occupied time here is useless. ```shell ./gg 16 24 0,1,2,3 run.sh ```