On November 12th, Inspur released its AI super-server AGX-5 at the SC18 International Conference for High Performance Computing, Networking, Storage and Analysis held in Dallas, USA. Capable of AI computing up to 2 petaflops per second within a single server, AGX-5 is one of the most powerful AI servers in the world. This year marks the 30th anniversary of the world’s largest supercomputing event, attracting more than 10,000 academic and industrial participators from around the world.
Inspur AGX-5 is designed to facilitate deep learning and high-performance computing. A single server carries 16 NVIDIA Tesla V100 Tensor Core 32GB GPUs with 10,240 Tensor Cores, providing up to 2 PFlops of performance. AGX-5 uses the industry’s most advanced NVIDIA NVSwitch interconnect fabric, based on NVIDIA’s latest HGX-2 platform, providing near linear AI performance acceleration. AGX-5 is also equipped with two 28-core CPUs that provide top-of-the-line general-purpose computing performance and 6 TB of persistent memory for high-speed data access.
Currently, more and more industries are working to integrate AI technology into their existing businesses, rapidly expanding the range of AI application scenarios. As the AI industry enters a major growth period, the speed and quality of AI innovation is becoming key to ensuring the continued competitiveness of many enterprises. Stronger AI computing platforms are needed. The AGX-5 is designed specifically to meet such needs.
“The capability of AI computing has become one of the key production factors that determine a company’s competitive edge and speed of innovation,” said Peter Peng, Vice President of Inspur Group. “NVIDIA is the world’s leading company in AI accelerated computing and a long-time partner of Inspur. As the latest product to come out of Inspur and NVIDIA’s collaboration, AGX-5 offers massively improved computing performance, chip interconnection, and data throughput. It will no doubt provide a major boost of computing power to help catalyze an unprecedented acceleration of AI development in commercial and research applications.”
“The value of fusing AI and HPC computing into a unified architecture based on NVIDIA Tensor Core GPUs is now recognized by more and more customers,” said Marc Hamilton, VP of Solutions Architecture and Engineering at NVIDIA. “Inspur has a deep understanding of customer demands and the ability to quickly, efficiently and innovatively develop scale-up computing systems based on the latest NVIDIA GPUs. The newly launched AGX-5 will help AI and HPC users worldwide break through their most complex computational bottlenecks and save significant cost, space, and energy in the data center.”
As the world’s leading AI computing provider, Inspur is fully engaged in the development of AI infrastructures on four layers, including computing platform, management & performance suite, optimized deep learning frameworks and application acceleration to deliver end-to-end, agile, cost-efficient, and optimized AI solutions for its industry customers. According to IDC’s 2017 China AI Infrastructure Market Survey Report, with 57% market share, Inspur ranks the first in the AI server market. Meanwhile, committed to offering a state-of-the-art computing edge for global customers through innovative design, Inspur has become a business partner of many leading companies in the world.