"NVLink will connect the machines’ processors – CPUs & GPUs – so they can exchange data 5 to 12 times faster. "
explanation (via .@nvidias tweet) : https://twitter.com/nvidia/status/534023085616148481
& additional article "How NVLink Will Enable Faster, Easier Multi-GPU Computing" on their blog
Now I'm kinda waiting when these technologies are going to hit consumer space =)
(at GTC'14 nVidia mentions that should happen with Pascal - maybe only on PRO line?)..
NVLink came to PRO line cards =)
You do not have the required permissions to view the files attached to this post.
- Seekerfinder
- Posts: 1600
- Joined: Tue Jan 04, 2011 11:34 am
That is interesting. They're very different architectures, so obviously Nvidia is designing it so that GPU apps can take advantage of it. But just quick look at the article suggests that CPU's will be able to access GPU processing power as well - or at least each other's memory. Is this the start of a unified architecture between CPU and GPU? If so it is a matter of time before one of the two architectures replaces the other completely in a single system... Fascinating stuff.
Seeker
Seeker
Win 8(64) | P9X79-E WS | i7-3930K | 32GB | GTX Titan & GTX 780Ti | SketchUP | Revit | Beta tester for Revit & Sketchup plugins for Octane
..we have not to forget that Intel is loosing a bit if nvidia is claiming the space on the highest bleeding edge with it's accelerators.
Intel also has Xeon Phi cards - curious what they are going to come with =) - interesting times to live with such technologies evolving =)
Intel also has Xeon Phi cards - curious what they are going to come with =) - interesting times to live with such technologies evolving =)