A preview is not available for this record, please engage by choosing from the available options ‘download’ or ‘view’ to engage with the material
Description
Preferred Networks (PFN) develops AI solutions for industrial and domestic robotics, IIoT, manufacturing systems and other industries. Their latest HPC system, MN-3, integrates a custom-designed deep learning accelerator they engineered. Traditional SSDs could not meet the I/O throughput requirements of the new architecture. PFN turned to Supermicro’s SuperServer hardware with 2nd Gen Intel® Xeon® Scalable processors and Intel® Optane™ persistent memory. This enabled a balanced node with fast access and high capacity for training data. PFN also developed code and a custom library to take advantage of the large capacity, low latency and byte-addressable features of Intel Optane PMem.