Inspur Announces AI Servers Supporting Latest NVIDIA A100 GPUs at SC20
Inspur’s AI servers NF5488A5 and NF5488M5-D, which provide support for the latest NVIDIA A100 80G GPU, are now available globally in mass production, delivering a 15% performance increase in AI model training with tens of billions of parameters.
San Jose, Calif. – November 16, 2020 – Inspur Information, a leading IT infrastructure solutions provider, today announced at Supercomputing Conference 2020 (SC20) that its AI servers NF5488M5-D and NF5488A5, which provide support for the latest NVIDIA A100 80G GPU, are now available globally in mass production.
The NF5488M5-D and NF5488A5 accommodate eight pieces of the next-generation A100 80G GPU with HBM2e memory, doubling the memory capacity and delivering a 30 percent increase in memory bandwidth of the GPU. The AI servers are designed to process larger amounts of data in a single node that could significantly reduce memory overhead in the communication between data or models in parallel, ultimately delivering a 15 percent performance increase in AI model training with tens of billions of parameters.
Inspur’s self-developed NF5488 is a next-generation series of AI servers that accommodate eight third-generation NVLink interconnected A100 GPUs in a 4U chassis, demonstrating a 234 percent performance increase compared with that of the previous generation. The NF5488 series provides users with leading AI computing power and mature ecosystem support for the most challenging AI applications, including intelligent customer service, financial analysis, smart cities, and natural language processing.
Recently, Inspur’s AI server NF5488A5 set 18 performance records in one fell swoop in the latest MLPerf Benchmarks, including the world record of 549,000 inferences per second for the ResNet50 inference task – a three-fold increase on the 2019 single-server performance record.
As the leading brand in AI computing, Inspur contains one of the most robust and extensive product portfolios in the industry. Inspur provides full-stack AI scenarios such as model training, machine reasoning, and edge computing, as well as the development of full-stack AI capabilities such as AI framework optimization, AI development management, and application optimization.