Based on the Caffe architecture, Inspur’s Caffe-MPI framework adopts MPI technology to maximize performance in data trainings through parallel data processing and multi-tasking and is supported on large-scale cluster platforms.
Inspur’s FPGA accelerator card F10A, based on Altera’s Arria 10 and leveraging OpenCL, substantially shortens software development time and offers high performance at low power consumption, essential to deep learning applications.
Teye analyzes AI application performance features of hardware and system resources on GPU clusters, providing analyses to help users maximize and optimize app computational potential.
Quad GPU in a 2U form factor supercomputer expansion box supporting multiple configurations
Video analytics AI server with 8 NVIDIA Tesla P4 GPUs / 4 FLHL GPUs and up to 96TB of local storage
Rack server with PCI-e Fabric interconnect architecture, ability to daisy-chain up to 64 GPUs