Optimized performance for modern workloads with NVIDIA-Certified Systems
The long-standing HPE and NVIDIA collaboration provides best-in-class compute hardware and software for optimal outcomes for accelerated AI workloads. Through their collaboration, HPE and NVIDIA deliver NVIDIA-Certified Systems that offer a unified architecture to accelerate traditional and modern workloads, from data center to edge. The single high-performance, cost-effective, and scalable infrastructure brings together compute and graphics acceleration and high-speed secure networking in the leading enterprise data center servers.
Differentiated hardware requires differentiated software
HPE and NVIDIA are perfect players in this hardware- and software-focused industry. HPE servers are certified NVIDIA systems that accelerate insights, delivering maximum ROI. NVIDIA offers an extensive collection of AI software to help users benefit from accelerated computing quickly and in a wide variety of areas, via the NVIDIA NGC catalog. This catalog is a GPU-optimized hub for AI, ML, and HPC containers, pre-trained models, SDKs, and Helm charts that simplify and accelerate end-to-end workflows. Data scientists and developers can make use of hundreds of containerized applications accelerated with NVIDIA technologies, allowing enterprises to immediately benefit from the NVIDIA-Certified systems.
The HPE Ezmeral software portfolio complements the NGC catalog by providing the basis for future state architecture for analytics and data intensive workloads. HPE Ezmeral Container Platform delivers the control plane to easily deploy and manage the NVIDIA NGC images on shared accelerated hardware. It allows clients to configure and reconfigure in minutes, based on the needs of the entire organization — not just one team at a time. Benefits include:
- Removing artificial silos by creating a shared pool of resources
- Replacing complex app and infrastructure deployment with containers
- Delivering on-demand, self-service GPU-accelerated analytics environments
Furthermore, the platform supports a fully open source and Kubernetes-based orchestration environment. This means there is a seamless integration of the HPE Ezmeral Container Platform that supports Kubernetes orchestration, allowing you to utilize resources either located on-premises, in the cloud, on the edge, or anywhere in between. With the HPE Ezmeral Container Platform, we have shown how silos of hardware, purpose-built for individual teams, can be eliminated. The combination of NVIDIA-Certified Systems from HPE, HPE Ezmeral Container Platform and ML Ops software, and the NVIDIA NGC catalog provides impressive benefits. Data scientists and developers can build applications faster, DevOps can deploy their applications reliably, and IT can support the entire process by deploying the ideal hardware configurations needed for the workloads.
Continuous improvement to get the most out of your data
According to Anant Chintamaneni, with the latest 5.3 release on April 5th, HPE Ezmeral Container Platform and ML Ops “makes it simpler than ever for clients to employ an industrialized approach to data science to support many users, tools, and steps in order to deliver an end-to-end, analytics solution to innovate faster.” This offering has a major focus on GPU-accelerated AI/Analytics and DevOps:
- Apache Spark™ on Kubernetes:
HPE Ezmeral now delivers enterprise on-premises Spark on K8s leveraging the open-source Apache Spark 3.X operator. This allows our clients to immediately take advantage of the latest and greatest Spark features like optimizations for GPU acceleration, GPU scheduling, and native data access from S3. Plus, we’ve added MLflow and Airflow integrations to improve collaboration and manage end-to-end data processing and data science pipelines.
- Maximize performance and utilization of heterogeneous GPU environments:
Now with support for the NVIDIA A100 Tensor Core GPU, enterprises get bare metal performance and maximum utilization with sharing of GPUs for data centric workloads across heterogeneous GPU types.
- Self-service data science:
We’ve made it even easier to bring your own tool (BYOT) with our App Workbench, which provides a guided workflow for Kubernetes application containerization. Using an enhanced GUI, you can create your own custom images or source from existing libraries, like NVIDIA NGC, in just a few clicks.
Learn how HPE and NVIDIA accelerate AI
To learn how HPE Ezmeral and NVIDIA advance AI through accelerated computing and GPU optimization:
- Watch GPU Optimization for Data Science with HPE Ezmeral – a short, two-minute video highlighting how HPE Ezmeral and NVIDIA accelerate AI deployments through accelerated computing and GPU optimization.
- Read the Press Release from Nanya Technologies on how they are using HPE Ezmeral and NVIDIA to accelerate the application of AI in its DRAM production.
- Attend NVIDIA’s largest developer conference, GPU Technology Conference (GTC) April 12-16th. Visit HPE at GTC to check out our sessions, register for the event, and register for a chance to win a $20 digital Starbucks gift card!
Matt Hausmann
Hewlett Packard Enterprise
twitter.com/HPE_Ezmeral
linkedin.com/showcase/hpe-ezmeral
hpe.com/HPE_Ezmeral
Matt_Hausmann
Over the past decades, Matt has had the privilege to collaborate with hundreds of companies and experts on ways to constantly improve how to turn data into insights. This continues to drive him as the ever-evolving analytics landscape enables organizations to continually make smarter, faster decisions.
