GPU accelerated computing offers unprecedented application performance by off-loading compute-intensive portions to the GPU. Each GPU consists of thousands of smaller and more efficient cores for handling multiple tasks. Applications are computed much faster in a parallel computing environment with GPU accelerating.

Our HPC cluster – Coral supports GPU accelerated computing with 12 NVIDIA® Tesla® V100 and 16 NVIDIA® GeForce GTX 1080 GPU. You can compile your own CUDA programs and then submit to the GPU computing nodes. CUDA-Aware MPI is also supported so that you can run a program across two GPU accelerated computing nodes for parallel computing.


  1. How do I compile a CUDA program in HPC?
  2. How do I run a CUDA program in HPC?
  3. How do I know my application supports GPU acceleration or not?
  4. What is CUDA-Aware MPI?