Product Requirements & Platform Availability for Parallel Computing Toolbox

System Requirements

• General resource requirements for parallel computing
  • Maximum of 1 MATLAB worker per physical CPU core is recommended.
  • Minimum of 2 GB RAM per MATLAB worker is recommended. If you are using Simulink, 4GB RAM per worker is recommended.
• Requirements for supporting the local scheduler
  • Minimum of 5 GB of disk space is recommended to accommodate temporary data directories.
  • If you are using Linux or Macintosh, and will create large parallel pools of workers, it is recommended that you adjust your operating system limits. For more information, see Recommended System Limits for Macintosh and Linux.
• Requirements for GPU Computing
  • CUDA-enabled NVIDIA GPUs with compute capability 3.0 or higher. For releases 17b and earlier, compute capability 2.0 is sufficient for all features except for use of deep learning. (GPU Support by Release Is my GPU supported?)
  • Latest graphics driver (Get the latest driver.)
• Requirements for scaling across multiple computers in a cluster or cloud
  • MATLAB Parallel Server extends the constructs of parallel computing to clusters and clouds.