News
News Categories

NVIDIA has announced two new Tesla P100-based deep learning initiatives with IBM and Microsoft

By Wong Chung Wee - on 15 Nov 2016, 1:43pm

NVIDIA has announced two new Tesla P100-based deep learning initiatives with IBM and Microsoft

The NVIDIA Tesla P100 GPU. (Image source:NVIDIA)

NVIDIA has announced two new Tesla P100-based deep learning initiatives with IBM and Microsoft. Both initiatives involve deep learning toolkits that leverages on the parallel compute prowess of the Pascal-based GPU.

The IBM Power System S822LC. (Image source: NVIDIA)

The new partnership between IBM and NVIDIA comprises hardware and development toolkits from both companies. Under this new initiative, IBM Power System S822LC will feature up to four NVIDIA Tesla P100 GPUs, which will be linked via NVLink architecture with a pair of IBM Power8 CPUs. In addition, IBM PowerAI toolkit will run in tandem with NVIDIA GPU deep learning toolkit; together, they will provide an AI platform that supports deep learning frameworks that include Caffe, DL4J, TensorFlow, Theano and Torch.

(Image source: NVIDIA)

Generally speaking, Microsoft isn’t a company that has been closely associated with high performance computing; however, it has joined forces with NVIDIA to optimize the Microsoft Cognitive Toolkit, formerly known as CNTK, for the NVIDIA DGX-1 deep learning supercomputer. The rebranded toolkit is also optimized for Azure N-Series Virtual Machines, and this means that corporations, who need to tap into trained data models, are able to tap into NVIDIA DGX-1 supercomputers on premise, or make use of Azure’s cloud computing services without having to invest in additional HPC hardware.

(Image source: NVIDIA)

The Microsoft Cognitive Toolkit has been the driving force behind popular Microsoft software and services like Cortana, Bing, Skype and even its augmented reality hardware, the HoloLens. In terms of performance gains when using the rebranded toolkit for image recognition training, NVIDIA touts an improvement of 170 times over a dual Intel Xeon E5 v4 CPU server! In short, the Microsoft Cognitive Toolkit is “supercharged” for the NVIDIA DGX-1 server node. In the near future with the help of the toolkit, companies will be able to leverage on NVIDIA GPUs hosted on Azure, and also on premise NVIDIA DGX-1 supercomputer, as “part of a hybrid cloud AI platform."

(Source: NVIDIA)

Join HWZ's Telegram channel here and catch all the latest tech news!
Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.