NVIDIA DGX Cloud Benchmarking

Optimize AI workload performance on NVIDIA AI infrastructure

Overview

Measure What Matters With DGX Cloud Benchmarking

NVIDIA DGX™ Cloud Benchmarking is a suite of tools, recipes, and services that take the guesswork out of measuring performance of AI workloads and infrastructure. As an extension of NVIDIA DGX Cloud, DGX Cloud Benchmarking provides a standardized and objective means of gauging performance across platforms, essential to optimizing AI workloads and speeding outcomes.

How Well Does Your Infrastructure Perform for the Latest AI Workloads?

Achieve higher performance of NVIDIA AI infrastructure and AI workloads with a suite of tools, recipes, and services.

Explore DGX Cloud Benchmarking Performance Recipes

Optimize your AI workload with performance recipes, and quickly set up and run standardized benchmarking methodologies in your environment.

What Is DGX Cloud Benchmarking?

Optimize AI workload performance on any NVIDIA infrastructure with DGX Cloud Benchmarking’s suite of tools, services, and recipes.

Using Performance Explorer, users can identify the ideal GPU count that minimizes both total training time and costs. The objective is to identify the right number of GPUs for a given workload that maximizes throughput and minimizes expenses—across projects and teams.

Features

Optimize Your AI Workloads With DGX Cloud Benchmarking

The Ingredients to Improve AI Workloads

Explore Performance Recipes, ready-to-use templates for evaluating the performance of specific AI workloads across hardware and software combinations.

Get the Most Out of Your AI Infrastructure

Use Performance Explorer to identify the right resources for a given workload to maximize throughput and minimize expenses—across projects and platforms.

Connect With Our Experts

Get guidance from NVIDIA advisors to benchmark and optimize your AI workloads when you join the DGX Cloud Benchmarking Program.

NVIDIA Exemplar Clouds

The NVIDIA Exemplar Clouds initiative raises the bar across security, usability, performance, and resiliency matching our DGX Cloud reference architecture

Benefits

Improve AI Workload Performance Across Platforms

Get the most out of your AI workload environments, and unlock the full potential of your AI infrastructure with DGX Cloud Benchmarking.

Measure True AI Workload TCO

Determine which platform can deliver the fastest time to train or desired GPU scale and at what cost using real-time and end-to-end performance data.

Accelerate AI Outcomes

Tune and optimize your AI workloads according to end-to-end metrics tailored to the performance of modern generative AI applications.

Benchmark Beyond the FLOPS

Evaluate beyond the GPUs, including infrastructure software, cloud platforms, and application configurations, to gain a holistic view of workload performance.

Align to a Definition of “Good”

Get a standardized and objective means of gauging platform performance, and understand the expected performance for given workloads or use cases.

Resources

Learn More About DGX Cloud Benchmarking

Next Steps

Ready to Get Started?

Supercharge AI workloads on NVIDIA AI infrastructure with DGX Cloud Benchmarking Performance Recipes.

Connect to Global GPU Compute

Discover, procure, develop, customize, and deploy AI applications and other GPU-accelerated workloads across multiple cloud providers.

Train, Tune, Deploy on DGX Cloud

Explore the fully managed platforms and services that let you bring mission-critical workloads from develop to deploy in the era of agentic and physical AI.