NVIDIA HPC Benchmarks

NVIDIA HPC-Benchmarks collection provides four benchmarks (HPL, HPL-MxP, HPCG, and STREAM) widely used in the HPC community optimized for performance on NVIDIA accelerated HPC systems.

NVIDIA’s HPL and HPL-MxP benchmarks provide software packages to solve a (random) dense linear system in double precision (64-bit) arithmetic and in mixed precision arithmetic using Tensor Cores, respectively, on distributed-memory computers equipped with NVIDIA GPUs, based on the Netlib HPL benchmark and HPL-MxP benchmark.

NVIDIA’s HPCG benchmark accelerates the High Performance Conjugate Gradients (HPCG) Benchmark. HPCG is a software package that performs a fixed number of multigrid preconditioned (using a symmetric Gauss-Seidel smoother) conjugate gradient (PCG) iterations using double precision (64-bit) floating point values.

NVIDIA’s STREAM benchmark is a simple synthetic benchmark program that measures sustainable memory bandwidth.

NVIDIA HPC Benchmarks package includes STREAM benchmarks optimized for NVIDIA GPU and NVIDIA Grace CPU.

System Support

This package is intended to support only:

• CPU Architectures: x86_64 and NVIDIA Grace CPU (Arm SBSA).

• GPU SM Architectures: NVIDIA Ampere GPU architecture (sm80), NVIDIA Hopper GPU architecture (sm90), and NVIDIA Blackwell GPU architecture (sm100).

• Linux distributions with glibc >= 2.28 – RHEL 8.8, SLES 15.5, and Ubuntu 22.04 and 24.04 have been tested.

• MPI libraries that are ABI-compatible with MPICH (e.g., MPICH, Cray MPICH, MVAPICH, etc.) and OpenMPI.

The x86_64 package includes only GPU benchmarks, while the Arm SBSA package offers both GPU and NVIDIA Grace CPU benchmarks.

Prerequisites

• CUDA 12.8 or newer

• OpenMPI 4.1 or newer, or MPICH 3.4 or newer

NVIDIA HPC Benchmarks structure

In addition to NVIDIA HPC benchmarks, the package contains:

• NVIDIA NCCL 2.25.1

• AWS OFI NCCL 1.6.0

• NVIDIA NVSHMEM 3.2.5

• NVIDIA GDR Copy 2.4

• NVIDIA NVPL BLAS 25.1 (Arm SBSA only)

• NVIDIA NVPL LAPACK 25.1 (Arm SBSA only)

• NVIDIA NVPL Sparse 25.1 (Arm SBSA only)

• LLVM OpenMP 18.1.1 (Arm SBSA only)

• TCMalloc 4.5.3 (Arm SBSA only)

Index

• Overview
  • Package structure
  • Using custom libraries
  • NVSHMEM
  • CUDA-aware MPI for Cray MPICH
• Release Notes
  • v25.04
  • v25.02
  • v24.09
  • v24.05
• NVIDIA HPL Benchmark
  • x86 package folder structure
  • aarch64 package folder structure
  • Running the NVIDIA HPL Benchmarks on x86_64 with NVIDIA GPUs, NVIDIA Grace Hopper, and NVIDIA Grace Blackwell systems
  • Running the NVIDIA HPL Benchmarks on NVIDIA Grace CPU only systems
  • References
• NVIDIA HPL-MxP Benchmark
  • x86 package folder structure
  • aarch64 package folder structure
  • Running the NVIDIA HPL-MxP Benchmarks on x86_64 with NVIDIA GPUs, NVIDIA Grace Hopper, and NVIDIA Grace Blackwell systems
  • Running the NVIDIA HPL-MxP Benchmarks on NVIDIA Grace CPU only systems
• NVIDIA HPCG Benchmark
  • x86 package folder structure
  • aarch64 package folder structure
  • Running the NVIDIA HPCG Benchmarks on x86_64 with NVIDIA GPUs
  • Running the NVIDIA HPCG Benchmarks on NVIDIA Grace Hopper, NVIDIA Grace Blackwell, and NVIDIA Grace CPU systems
• NVIDIA STREAM Benchmark
  • x86 package folder structure
  • aarch64 package folder structure
  • Running the NVIDIA STREAM Benchmarks on x86_64 with NVIDIA GPUs, NVIDIA Grace Hopper, and NVIDIA Grace Blackwell systems
  • Running the NVIDIA STREAM Benchmarks on NVIDIA Grace CPU only systems
• Software License Agreement

Support

For questions or to provide feedback, please contact HPCBenchmarks@nvidia.com

Resources

• World’s Fastest Supercomputer Triples Performance Record | NVIDIA Blog

• Netlib HPL benchmark

• HPL Mixed-Precision Benchmark

• HPCG Benchmark

• Developer Forums

• NGC NVIDIA HPC-Benchmarks container

• NVIDA HPCG

• Accelerating the HPCG Benchmark with NVIDIA Math Sparse Libraries