AI acceleration with NVIDIA and RED HAT

NVIDIA AI Enterprise is an end-to-end, cloud-native suite of AI and data analytics software optimized, certified, and supported by NVIDIA to run on Red Hat OpenShift, the industry-leading Kubernetes platform, with NVIDIA-certified systems. The solution incorporates key technologies from NVIDIA and Red Hat to securely and consistently deploy, manage, and scale AI workloads in the hybrid cloud.

Read the solution brief to learn more about NVIDIA AI Enterprise on Red Hat OpenShift.

Red Hat® OpenShift® is an enterprise-ready Kubernetes container platform designed for an open hybrid cloud strategy. It leverages the NVIDIA® BlueField DPU® to offload the server CPU, accelerating infrastructure tasks for faster, more efficient, and more secure servers and networks. Together, Red Hat® OpenShift® and NVIDIA BlueField® DPU provide a consistent, cloud-native application platform for managing hybrid cloud, multi-cloud, and edge deployments with improved orchestration, automation, and security.

NVIDIA and Red Hat offer an accelerated AI infrastructure for on-premise deployments.

Built on Red Hat OpenShift, the joint solution is a platform to accelerate ML/DL modeling for data scientists and deliver intelligent AI-powered applications.

Enterprises can leverage NVIDIA's NGC™ software hub to access GPU-optimized DL and ML applications and pre-trained AI models to take advantage of massive parallel processing. Leading server vendors recommend NVIDIA-certified systems with NVIDIA GPUs validated to run accelerated applications with optimal performance.

The NVIDIA GPU Operator for Red Hat OpenShift simplifies and accelerates compute-intensive ML/DL modeling tasks for data scientists and inference tasks in data centers, edge and public clouds. The GPU Operator enables Red Hat OpenShift to schedule workloads that require NVIDIA GPUs as easily as scheduling CPU or memory for traditional workloads. Typical use cases that can benefit from GPU acceleration include image and speech recognition, conversational AI, and visual search.

The joint solution is compatible with NVIDIA certified systems and Red Hat Enterprise Linux® certified servers.

When building modern data centers, it is essential that hardware and software go hand-in-hand. The collaboration of NVIDIA® and Red Hat® enables the delivery of accelerated networking solutions that create a robust, scalable, and secure infrastructure for modern workloads. This can help you deploy an optimized software-defined infrastructure, accelerate it, and develop applications in less time, leading to improved operational efficiency and competitive differentiation.

Providing a standardized and accessible infrastructure for enterprise digital transformation is a must. The collaboration between the two companies helps you deliver networking hardware innovations for cloud-native computing through Red Hat OpenShift and cloud networking through OpenStack. By leveraging NVIDIA technologies such as GPUDirect RDMA, Accelerated Switching and Packet Processing (ASAP2), and the Morpheus AI Cybersecurity Framework, enterprises are ready for the next wave of applications.

• highest energy efficiency
• highest area efficiency
• highest throughput
• fastest warm and cold start times in the large Greeks benchmark
• fastest warm-up time in Greeks benchmark
• highest maximum paths
• highest maximum assets

Below is an overview chart of the benchmark configuration that was used.