Setting a new bar for AI workload performance across clouds.
Overview
NVIDIA Exemplar Cloud seeks to improve performance per TCO for all cloud providers with hardware and software recipes, references, tools, and capabilities that solve infrastructure challenges across key metrics, including workload performance, security, and reliability. The initiative leverages NVIDIA DGX™ Cloud Benchmarking recipes to establish standardized benchmarks for AI workload performance across the cloud provider ecosystem. This ensures consistent, high-quality GPU performance for end users, helping to drive consumption, improve marketability, and strengthen NVIDIA’s relationships with cloud providers.
Benefits
Exemplar Cloud helps customers align to transparent and standardized benchmarking, so they can compare cloud AI infrastructure based on real workload performance and total cost of ownership.
Cloud providers can work together with NVIDIA to assess workload performance, debug hardware, software, and platform issues, and implement improvements.
Providers can use Exemplar Cloud to drive highly tuned and optimized cloud infrastructure, differentiating themselves with superior AI workload performance across training, fine-tuning, and inferencing.
With Exemplar Cloud, NVIDIA and the provider community can drive advancements in platform consistency, TCO optimization, and customer trust by making performance results public and building a transparent, trusted ecosystem.
Partners
Exemplar Cloud providers have collaborated with NVIDIA to optimize their infrastructure to ensure their customers’ workloads will be performant, secure, and reliable, based on real-world workload and TCO metrics. Cloud providers who achieve the metrics-based threshold can become an NVIDIA Exemplar Cloud.
Next Steps
Unlock AI benchmarking insights by connecting with NVIDIA to optimize and accelerate your projects.
Explore ready-to-use templates for evaluating performance of specific AI use cases across hardware and software combinations.
Learn more about NVIDIA DGX Cloud Benchmarking, a suite of tools for optimizing AI workloads on various platforms.