The new POWER9-based AC922 Power Systems are the first to embed technologies such as PCI-Express 4.0, and the next-generation NVIDIA NVlink 2.0 and OpenCAPI which combined can accelerate data movement, calculated at 9.5x faster than PCI-E 3.0 based x86 systems. This can drive demonstrable performance improvements across popular AI frameworks such as Chainer, TensorFlow and Caffe, as well as accelerated databases such as Kinetica.
Complimenting these IO technologies are advanced accelerators, including up to six NVIDIA Volta-based Tesla V100 GPUs, Xilinx FPGAs, and Mellanox Infiniband adaptors to drive unprecedented performance on today’s most-demanding workloads like AI.
The massive data throughput and processing power due to the latest interconnects and GPU capabilities allow customers to more quickly deliver business value. Combined with PowerAI, IBM’s industry-leading deep learning toolkit, the AC922 will deliver near linear scaling for distributed deep learning workloads for scale out AI.
For more information please contact XENON at firstname.lastname@example.org or 1300 888030.