Added: 1 year ago
From: Scobleizer
Views: 2,437
Sort by time | Sort by thread (beta)

Link to this comment:

Share to:

All Comments (11)

Sign In or Sign Up now to post a comment!
  • It will be interesting to see the comparative metrics :with and w/o GPU on their setup. Really cool though that this thought process has started catching hold i the web-infra space and guys like Jack adopt this approach/lead the charge. Especially as the real-time constraints on analytics increase and the data-sets increase pure MW optimization wont cut it and kernel level optimization (scheduler mods) likely wont give you the ROI or scalability or TOC.

  • Glad atleast some folks in the space are thinking out of the box.

  • Nothing new here. Pitched something like this ,but with FPGAs (programming paradigm is different) to Amazon about 4 years ago, for their EC cloud. No buy in from them. Usually the APP-SW guys are a generation behind in adoption and thought process. GPU are power intensive and have a shared memory bottleneck, no ideal for Datacenter. We will let the App SW guys find that out in another 4 years!! He is very right, it is not hard to do. Glad to see that the APP guys are getting tech savvy...

  • Paper of the work he's talking about?

  • It's really amazing ... You can not only save space, cost but also a huge amount of energy by NVIDIA GPUs.

  • Pics or it didn't happen !

  • Awesome

  • This is sooooo cool!! Thanks.

  • Blown away!!

  • Wow! Truly I'm

  • /headexplodes

    An elegant idea, which means it is both brilliant and simple.

Loading...
Alert icon
0 / 00Unsaved Playlist Return to active list
    1. Your queue is empty. Add videos to your queue using this button:
      or sign in to load a different list.
    Loading...Loading...Saving...
    • Clear all videos from this list
    • Learn more