NVIDIA Topology-Aware GPU Selection (NVTAGS) is a toolset for high-performance computing (HPC) applications that enables faster solve times with high GPU communication-to-application run-time ratios. NVTAGS intelligently and automatically assigns GPUs to message passing interface (MPI) processes, thereby reducing overall GPU-to-GPU communication time.

Get started


NVTAGS dramatically improves performance of single- and multi-node workloads by intelligently mapping MPI processes to GPUs for HPC applications with non-uniform GPU-to-GPU communication.


It provides a fully automated process compatible with Slurm, Singularity, and bare-metal runs that eliminates the manual, time-consuming mapping of processes to GPUs.

Seamless Integration

It’s extremely portable and can be easily integrated into your Slurm workflow or added to an application container, making it easily deployable on any system or architecture.

Key Features

  • Automated topology detection and CPU and network interface card (NIC) binding, independent of the system and HPC application
  • Support for single- and multi-node, PCIe, and NVIDIA® NVLink® with NVIDIA Pascal™, Volta™, and Ampere architecture GPUs
  • Straightforward integration with Slurm and Singularity
  • Small memory footprint (few KB) with minimal one-time profiling overhead
  • Automatic caching of mapped results for future simulations
  • Support for customizing default mapping (e.g., mapping algorithm) and profiling metrics

Performance Gains

NVTAGS Multi-Node Performance Improvements for Chroma, MILC and LAMMPS

Performance highlights:

  • Up to 27% performance increase for Chroma
  • Up to 75% performance increase for MILC
  • Up to 64% performance increase for LAMMPS


Ready to get started with NVTAGS?

Get Started