What can I help you with?
NVIDIA Messaging Accelerator (VMA) Documentation Rev 9.8.60

System Requirements and Interoperability

The following table presents the currently certified combinations of stacks and platforms, and supported CPU architectures for the current VMA version.

Specification

Value

Network Adapter Cards

NVIDIA ConnectX-7 (4x25G)

NVIDIA ConnectX-7 (200G) Crypto-disabled

NVIDIA ConnectX-5 / NVIDIA ConnectX-5 Ex

Firmware

ConnectX-7 28.44.1036

ConnectX-5 16.35.3502

Driver Stack

  • DOCA-Host v2.10.0

    • doca-libvma profile

    • doca-roce profile

NOTE : Starting this version, the host driver is part of the NVIDIA DOCA package.

For further information, please see NVIDIA MLNX_OFED to DOCA-OFED Transition Guide .

Supported Operating Systems and Kernels

  • Ubuntu 20.04

  • RedHat 8.6, 8.7, and 9.3 ( for NVIDIA GH200 Grace Hopper only )

CPU Architecture

  • x86_64 (Intel Xeon)

  • Arm GH200 Grace Hopper

Minimum memory requirements

1 GB of free memory for installation 800 MB per process running with VMA

Minimum disk space requirements

1 GB

Transport

Ethernet/InfiniBand

Description

Binary RPM and DEB packages for 64-bit architecture for Linux distribution

  • libvma-9.8.60-1.x86_64.rpm

  • libvma-devel-9.8.60-1.x86_64.rpm

  • libvma-utils-9.8.60-1.x86_64.rpm

  • libvma_9.8.60-1_amd64.deb

  • libvma-dev_9.8.60-1_amd64.deb

  • libvma-utils_9.8.60-1_amd64.deb

Documentation

VMA Release Notes

VMA Installation and Quick Start Guide

VMA User Manual

The VMA library has been successfully tested and is certified to work with the applications listed in the following table.

Application

Company / Source

Type

Notes

sockperf

NVIDIA® (Open Source)

Bandwidth and Latency Benchmarking

Version 3.10 (https://github.com/mellanox/sockperf)

netperf

Open Source

Bandwidth and Latency Benchmarking

Version 2.6.0

NetPIPE

Open Source

Network Protocol Independent Performance Evaluator

Version 3.7.2

UMS (formerly LBM)

Informatica

Message Middleware Infrastructures

Version 6.7

© Copyright 2025, NVIDIA. Last updated on Dec 2, 2024.