Nvidia Releases Tesla P100 GPU
A big player has finally made its way to the world of supercomputers.
It was stated that Nvidia has released the Tesla p100 GPU, which is expected to be the fastest GPU for a supercomputer. The design of the GPU is influenced from Pascal, which gives it a new standard of instructions for scientific computations and learnings. Nvidia announced that there will be 4,500 GPU inputted into the supercomputer called Piz Daint.
Reports also showed that the Piz Daint will be able to perform with 7.8 teraflops. This speed will put it at seventh place for the faster computer in the world. In addition, the new supercomputer will be utilized at CERN to analyze data from the Large Hardon Collider.
Get Our Latest News for FREE
However, aside from the Piz Daint, the developers of Tesla p100 GPU have also created a desktop supercomputer that uses the Telsa p100, called DGX-1. The device is able to perform with 170 teraflops, which puts it in the competition for one of the fastest computers in the world. The DGX-1 also has two Xeon CPUs and 7 tera bytes worth of solid state drive storage. The device also has dual 10 gigabit ethernet ports.
The new GPU is also going to be used in the IBM servers of Hewlett Packard Enterprises in the coming year. It will help in developing the big data centers that need the cloud servers since the GPU has one of the largest and fastest chips made.
Aside from the capability of its chips, the programming framework will allow developers to create applications for the computer. The Tesla will also have 16 giga bytes of high bandwith memory 2, which is the latest type of memory available for GPUs. It allows for the chips to be stacked on top of each other as opposed to side by side. It will give it a faster and denser type of memory.
