Advancing AI Capabilities with Next-Generation HPC Solutions

In this evolving digital economy, data is the cornerstone of success. Big Data is redefining the way we think, act, and understand the world, and accelerating insight is the difference between making the next major discovery and missing it. The more information we can effectively capture, analyze, and act on, the more opportunities there are to drive technological advancements, ensure economic control, strengthen national security, and fuel scientific research.

Organizations across all sectors are putting this data to work. They are optimizing their IT operations and enhancing the way they communicate, learn, and grow their businesses in order to harness the full power of artificial intelligence (AI). Backed by high performance computing (HPC) technologies, AI is revolutionizing the world as we know it—from web searches, digital assistants, and translations; to diagnosing and treating diseases; to powering breakthroughs in agriculture, manufacturing, and electronic design automation.

However, machines have already exceeded human levels of performance, and organizations are quickly approaching the peak of their IT capabilities. Today’s systems are struggling to keep pace with increasing demands, and future workloads will require greater memory capacity and higher bandwidth to operate efficiently. In order to continuously innovate and evolve, organizations are seeking the next generation of HPC solutions to achieve superior productivity, speed data-driven decision-making, and facilitate growth.


The demand for faster, smarter, and more reliable compute platforms is rising exponentially as IT departments strive to keep pace with proliferating data volumes and derive Deep Learning insights. Today’s IT leaders are exploring solutions with superhuman intelligence to harness the full power of AI and pioneer the next generation of HPC systems.

To tackle increasingly diverse challenges, neural network complexity is exploding, leveraging HPC tools and advanced services, such as Deep Learning and GPU cloud computing. Introduced in 2015, Microsoft ResNet reached 7 exaflops with 60 million parameters to power superhuman image recognition. In 2016, Baidu Deep Speech 2 harnessed 20 exaflops with 300 million parameters to enable superhuman voice recognition. Neural Machine Translation released by Google earlier this year achieved 100 exaflops with 8700 million parameters for near-human language translation.

Topping this list of solutions, the NVIDIA Volta is fueling some of the most powerful supercomputers in the U.S. By combining AI with traditional HPC applications on a single platform, Volta rapidly accelerates workloads for HPC, AI training, AI inference, and virtual desktops. Powered by Volta, the NVIDIA Tesla V100 pairs 5,120 CUDA cores and 640 NEW tensor cores to deliver 120 TeraFLOPS of Deep Learning, 7.5 TeraFLOPS of double precision performance, and 15 TeraFLOPS of single precision performance to turbocharge both HPC and AI, making it the most advanced data center GPU ever built. V100 delivers 3X faster Deep Learning performance and 1.5X higher HPC performance over Pascal GPUs. This includes delivering over 80X Deep Learning training performance in 3 years, 3X reduction in training time over the NVIDIA Tesla P100, and efficient scaling to multiple GPU nodes.

In addition to enhancing speed and accessibility, the NVIDIA TensorRT, a Deep Learning inference optimizer and runtime engine, enables 3.5X faster inference performance and delivers dramatic throughput gains, even at less than 7 milliseconds of latency required by real-time AI services. This solution works to rapidly optimize, validate, and deploy trained neural networks to hyperscale data centers. Now, developers can easily optimize their AI-powered applications for even the most demanding throughput and latency requirements. TensorRT, optimized for Volta, will be available later this year as a free download for members of the NVIDIA Developer Program.


Hewlett Packard Enterprise (HPE) is collaborating with NVIDIA to revolutionize AI and Deep Learning capabilities. Future HPE systems are expressly designed to support the Tesla V100 card, including HPE Apollo, HPE ProLiant DL, and HPE Integrity MC990 X systems. HPE is excited to complement our purpose-built systems innovation for Deep Learning and AI with the unique, industry leading strengths of the NVIDIA V100 technology architecture to accelerate insights and intelligence for our customers. HPE will support NVIDIA Volta with PCIe interconnects in three different systems in our portfolio and provide early access to the next generation of NVLink systems for select customers to address emerging customer demand.

Together, HPE and NVIDIA are industrializing AI and Deep Learning, empowering organizations with extreme versatility, optimal performance, and reduced time-to-insight. These HPC innovations promise to dramatically increase our knowledge and IT capabilities and achieve untapped levels of Compute—all the way to exascale.

To learn more about this powerful collaboration as well as our critical next steps in the race to the future of HPC, I encourage you to follow us on Twitter @HPE_HPC and NVIDIA @NVIDIA. You can also follow me at @VineethRam for up-to-the-minute news and updates. mitted if need be.

Sign up to our Newsletter

Featuring highlights, analysis, and stories from the week directly from us to your inbox with nothing in between.
Subscribe now