Comments on: DGX-1 Is Nvidia’s Deep Learning System For Newbies https://www.nextplatform.com/2016/04/06/dgx-1-nvidias-deep-learning-system-newbies/ In-depth coverage of high-end computing at large enterprises, supercomputing centers, hyperscale data centers, and public clouds. Mon, 23 Apr 2018 11:16:03 +0000 hourly 1 https://wordpress.org/?v=6.5.5 By: Sharpy https://www.nextplatform.com/2016/04/06/dgx-1-nvidias-deep-learning-system-newbies/#comment-56484 Sat, 23 Jul 2016 10:34:31 +0000 http://www.nextplatform.com/?p=2992#comment-56484 What software price are you talking about? Ubuntu, Cuda drivers and free deep learning framework? It’s $0.

]]> By: OranjeeGeneral https://www.nextplatform.com/2016/04/06/dgx-1-nvidias-deep-learning-system-newbies/#comment-42272 Wed, 06 Apr 2016 17:39:27 +0000 http://www.nextplatform.com/?p=2992#comment-42272 In reply to Igor.

Totally agree with you. I think that whole Pascal GPU is a bit of a smokescreen, there seem to be no specific architectural changes that make it better for these specific tasks. HBM2 maybe but that is pretty general for all kind of compute problems. The race to lower bit structures now we’re seeing 1-Bit Neural networks via XNORnets will make the whole GPU infrastructure completely obsolete. As Integer performance always has been rather lacklustre on GPUs

]]> By: Igor https://www.nextplatform.com/2016/04/06/dgx-1-nvidias-deep-learning-system-newbies/#comment-42267 Wed, 06 Apr 2016 15:25:26 +0000 http://www.nextplatform.com/?p=2992#comment-42267 At this price level I’m very doubtful DGX-1 has any price/performance advantage whatsoever over easily and immediately available from Supermicro, Tyan etc. servers with 8 Tesla K80 GPUs (with possible exception of apps/kernels that can use FP16 data instead of FP32).

]]>