Chapter 20 Accelerators

Moab can integrate with the Torque resource manager to discover, report, schedule, and submit workload to various accelerator architectures (such as NVIDIA GPUs or Intel® Xeon Phi™ co-processor architecture) for parallel processing. See the topics below for specific information.

• Scheduling GPUs

  • Using GPUs with NUMA
  • NVIDIA GPUs
  • GPU Metrics
• Configuring Intel® Xeon Phi™ Co-processor Architecture
  • Intel® Xeon Phi™ Co-processor Metrics

© 2017 Adaptive Computing