Intel is working on discrete GPUs
First devices are expected by 2020
First devices are expected by 2020
Oak Ridge National Laboratory’s latest machine features chips from IBM and Nvidia
16 Tesla V100 GPUs power this beast of a system, set to be manufactured by major server vendors
Plus Kaldi and Kubernetes optimization and acceleration come out of Nvidia’s annual show
And they’re going to use it for their in-house supercomputer
But they’re not selling it - yet
More bang for your buck
AI continues to push data center chip deployments
Instead, it wants you to buy the significantly more expensive Tesla
Academics from eight universities and IT researchers will have access to the system
Country’s AI Grand Plan enlists the GPU king
Unless China reveals a surprise supercomputer
Joins a growing number of Chinese corporations in offering GPUs for AI applications
A hybrid machine-learning supercomputer, based on SGI, Nvidia and DDN equipment
And they couldn’t have done it without GPU
Bets on neural nets
There’s also a new HPC server to go with it
Nvidia corners GPU accelerator market
Latest AI supercomputer is like a 250-server data center in a 3U box
The company wants to democratize access to cutting-edge technology
Grid 2.0 GPU infrastructure improves on the original