IBM/NVIDIA Launch Linux-Based Servers for Machine Learning

International Business Machines Corporation IBM has launched a new lineup of Linux-based servers that it built in collaboration with NVIDIA Corporation NVDA . The servers have been specifically designed to speed up artificial intelligence, advanced analytics and machine learning as well as to boost the efficiency of data centers. The new lineup of servers is powered by two IBM Power8 CPUs along with four Tesla P100 GPU accelerators manufactured by NVIDIA. The reason behind the high speed performance of these servers is that they also use NVIDIA's NVLink high-speed interface, which acts as a bridge between the CPU and GPU. This enables faster communication than a common PCIe bus found on a desktop computer.