Today’s enterprises need deep learning, but most don’t know how to get started. As rising data volumes and evolving industry trends push the limits of traditional IT, the latest innovations are helping them operate faster and smarter—and high performance computing is just the beginning.
Enterprises are deploying robust server platforms to power HPC applications, leveraging optimal performance, reliability, and flexibility to handle increasingly dense workloads. And with these industry-leading tools, modeling and simulation capabilities are rapidly evolving. Artificial intelligence is transforming how we operate and relate to technology. AI allows machines to think and learn like the human brain, while reducing time to intelligence and improving the accuracy of results. Some organizations are already utilizing AI algorithms to streamline search engines and power virtual assistants (like Siri, Alexa, and Cortana), but these capabilities only scratch the surface.
Deep learning, a powerful subset of AI, is driving profound advancements in a variety of disciplines such as genomics sequencing analytics, climate research, and self-driving cars. AI and deep learning techniques are rapidly expanding the scope of human knowledge. Yet many organizations have immature, sub-optimal infrastructure, limited deep learning expertise and resources, or lack of technology integration capabilities needed to meet the extreme performance requirements of deep learning. In order to harness the full power of AI, IT departments must create an IT environment with solutions that are easy to deploy and manage, highly efficient, integrated, and scalable to handle the challenges of tomorrow.
Accelerating Intelligence With Deep Learning
Hewlett Packard Enterprise is offering several new capabilities to help users simplify AI usage and demystify deep learning. These include expertise and integrated solutions to get started, proven blueprints and services to scale and integrate, as well as technology integration to optimize deep learning environments.
HPE is helping customers to rapidly develop deep learning models with simple, “get-started” deep learning solutions and state-of-the-art training:
- HPE Deep Learning Cookbook – The HPE Deep Learning Cookbook provides an easy-to-follow “recipe” to implement deep learning tools. This technology configuration guide helps enterprises estimate and refine performance, characterize frameworks, and select the ideal hardware and software stacks to fit their individual needs.
- HPE Deep Learning Configuration for Rapid Software Installation – For easy setup and choice of OS, HPE offers an integrated solution with HPE Apollo 6500 and Bright Computing. With Bright Cluster Manager, an OS provisioning and software installation tool for clusters, you can install the Linux operating system of choice across the cluster. Further you can install the workload management tool of choice (PBS Professional, SLURM, or Moab).
- Enhanced Education and Training – HPE and NVIDIA are providing training for developers, data scientists, and researchers. At the NVIDIA Deep Learning Institute (DLI), customers have access to online labs and hands-on workshops that will prepare their workforces to utilize the latest deep learning techniques and software frameworks.
HPE is offering tools and services to help scale and integrate production environments to deliver value from deep learning investments. With new integrated hardware and software with automatic updates, these solutions are designed to scale and dramatically increase productivity. Enterprises stand to gain a great deal from these capabilities:
- Improved Deployment Effectiveness – To streamline deep learning applications, HPE has developed a new reference design for Advanced Image Classification. This trains neural networks to identify objects contained in images quickly, accelerating informed decision making, eliminating the guesswork, and simplifying deep learning integration.
- Integrated Production Environment – To further assist customers, HPE experts in the HPE AI Innovation Labs help run performance benchmarking remotely for a variety of use cases, algorithms, and frameworks. Customers can develop reference architectures and configurations to enable deep learning platforms and solutions as well as validate and develop relevant software for quick, reliable deployments.
- Financial Flexibility – With HPE Flexible Capacity, users can realize the freedom of public cloud with the security of on-premises IT, paying as they go. A flexible payment model helps enterprises avoid over-provisioning, increase cost savings, and scale as needed to accommodate business growth.
Technology Integration Capabilities To Maximize Performance
For customers who have mature production deep learning environments and want to seek additional competitive advantage, HPE offers advanced capabilities and expertise to maximize their performance. Customers can boost their application performance by integrating the latest Nvidia GPU technologies and glean expert advice to accelerate time to value in our HPE and Nvidia Global Centers of Excellence (CoEs):
- Latest Nvidia GPU-Accelerated Platforms – HPE is pleased to announce support for the Nvidia Tesla V100 PCIe cards on our HPE Apollo 6500 and HPE ProLiant DL380 Nvidia Tesla V100 GPUs are the world’s most advanced data center GPUs, delivering a robust platform for HPC and AI convergence, architected from the ground up to simplify programmability and improve GPU utilization by sharing resources among small jobs.
- Enhanced HPE and Nvidia CoEs – For customers wanting to accelerate their time to value, HPE and Nvidia have strengthened their CoE capabilities, providing access to the latest technologies, solutions, and expertise to optimize their deep learning environments. Services include technology and investment planning, code modernization, performance benchmarking, proofs of concept, and more.
Innovating For The Future
Cognitive computing is driving major scientific, technological, and economic advancements all over the world. On September 27 and 28, business leaders from a variety of industries came together in San Francisco to discuss the next steps in AI innovation. The AI Summit is the world’s first and largest conference/exhibition, geared toward bringing AI capabilities to corporate organizations. The Summit featured more than 50 sponsors and exhibitors, more than 60 sessions, over 80 press and media, more than 120 speakers, over 2,000 delegates, and in excess of 12,000 networking meetings.
HPE showcased several innovative and unique capabilities in our booth. New offerings include HPE support for Nvidia Tesla V100 GPUs, the industry’s most comprehensive compute portfolio for deep learning (based on select HPE Apollo systems), the HPE Deep Learning Cookbook, a new deep learning infrastructure bundle, new deep learning reference configurations for image classification, new HPE Pointnext Services, and new validated deep learning reference architectures. Our exciting innovations cover the expansion of our deep learning partner ecosystem, enhanced Global CoEs for deep learning, and advanced education and training services.
HPE is striving to demystify deep learning, making AI-driven tools widely available and deployable for rapid, actionable intelligence. We are helping customers rapidly develop deep learning models, scale and integrate, and optimize their IT to reach breakneck levels of performance and groundbreaking insight. Together, HPE and Nvidia are helping organizations evolve and grow, providing the necessary tools to harness next-generation AI capabilities.
Pankaj Goyal is vice president of the AI Business at Hewlett Packard Enterprise.