HPE: Your strategic partner for AI

JUNE 20, 2023 • BLOG POST • NEIL MACDONALD, EXECUTIVE VICE PRESIDENT & GENERAL MANAGER, COMPUTE

IN THIS ARTICLE

  • Today we’re launching three new AI inferencing solutions to speed your time to value. These systems are certified and optimized for inference at the edge and in the data center.
  • With these announcements at HPE Discover, HPE has changed the conversation about AI from “how to” to “how fast.”

HPE is changing the conversation about AI from “how to” to “how fast.”

AI is the “killer app” of this decade. There remain, however, significant hurdles to tap into the promise of this new transformational generation of AI models; training large language models at scale can be prohibitively expensive and stitching together solutions can be time consuming and requires deep expertise.

That’s all changed now with a series of announcements we’re making this week at HPE Discover. HPE is taking an end to end approach to AI so you can unlock the potential of AI for your business. Whether you need to deploy in the cloud, the data center or at the edge, we have you covered. Across training, tuning or inference, regardless if your business demands results in days, hours, or seconds, HPE has a solution that fits your needs.

Up first, we announced a new AI cloud portfolio, the first of which is HPE GreenLake for Large Language Models. This solution combines HPE’s AI software with HPE’s market-leading supercomputers to allow enterprises to train and tune large-scale AI models on demand, in a multi-tenant cloud service – privately and sustainably. Unlike public cloud offerings that run multiple workloads in parallel, HPE GreenLake for Large Language Models is architected to run a single large-scale AI training and simulation, and at full computing capacity. The offering will support AI and HPC jobs on hundreds or thousands of CPUs or GPUs at once.  This capability is significantly more effective, reliable, and efficient to train AI and create more accurate models, allowing enterprises to speed up their journey from POC to production to solve problems faster.

Second, we’re bringing to market three new AI inferencing solutions to speed your time to value. These systems are certified and optimized for inference at the edge and in the data center. The solutions are built on HPE’s new ProLiant Gen11 servers and integrate NVIDIA accelerated computing. HPE ProLiant Gen11 servers are designed from the ground up to support advanced accelerators. These new HPE ProLiant AI solutions boost inference performance by more than 5X* from previous generation systems.

Ideal for computer vision and video analytics, the HPE ProLiant DL320 Gen11 has a unique compact design, purpose-built for edge computing. It can pack up to four NVIDIA L4 Tensor Core GPUs in a 1U form factor to power smart spaces and loss prevention solutions and deliver insights in near real time. Customers can take advantage of offerings from the NVIDIA Metropolis ecosystem to deploy solutions targeted for their industry and use case.

For Generative Visual AI to drive product design, 3D animation or image and video generation, the HPE ProLiant DL380a Gen11 with up to four NVIDIA L40 GPUs delivers the rendering and design performance needed by demanding visual applications. The NVIDIA AI Enterprise software offers enterprise-grade support, security and stability for more than 100 pretrained models, frameworks and tools that streamline development and deployment of AI.

And finally, to drive enterprise implementations in AI natural language processing and inferencing applications like speech AI and fraud detection, HPE designed the ultra-scalable HPE ProLiant DL380a Gen11. Together with NVIDIA H100 Tensor Core GPUs and the full suite of NVIDIA AI Enterprise software, including NVIDIA Riva, this is a powerful AI platform.

Our solutions extend past compute. When I said we have an end to end offer, that includes the experts on staff to guide you from day zero with our HPE AI Transformation Workshop and the software you need to develop and deploy your models; HPE Machine Learning Data Management Software and HPE Machine Learning Development Environment. 

With these announcements at HPE Discover, HPE has changed the conversation about AI from “how to” to “how fast.” HPE delivers the full life cycle of AI development and deployment, from training the largest models in the cloud to inferencing at the edge in near real-time, enabling you to harness the transformative power of AI for your business.
 

* NVIDIA: : Comparison for image generative AI performance of NVIDIA L40 (TensorRT 8.6.0) versus T4 (TensorRT 8.5.2), Stable diffusion v2.1 (512x512)

RELATED NEWS