Alert icon
We're changing our privacy policy. This stuff matters.  Learn more  Dismiss

OpenCL for CPU Scales Applications Easily on Four Six-Core AMD Opteron Processors

Loading...

Sign in or sign up now!
51,777
Loading...
Alert icon
Sign in or sign up now!
Alert icon

Uploaded by on Aug 3, 2009

AMD demos its OpenCL for CPU beta on Six-Core AMD Opteron Processors. See how easily the Powder Toy physics simulation application scales across 24-cores with OpenCL without making changes to the application.

  • likes, 4 dislikes

Link to this comment:

Share to:

Top Comments

  • I'm looking at building a system with 48 cores and 500 Gigs of ram using the SAS as it was meant to be used with my SSD hard drives working at a GIG/s off the PCi express bus. Possible yes. Why because I need a Real mans Calculator.

  • Will it blend?

see all

All Comments (75)

Sign In or Sign Up now to post a comment!
  • Hello, can I ask you, is this CUDA open CL system avaiable to count CFD computation of fluid dynamics?? could you please write me ate richard.pinka@gmail.com ?, thank you

  • Why wouldn't you have a repeatable scenario to show a true comparison? And show some actual numbers?

  • name "game" is - The Powder TOY - ;)

  • Wish they'd hurry up and tell us how the hell we're meant to get openCL to work in TPT! Fed up of having an 8 threaded i7 and still lagging on big sims.

  • Nvidia is way ahead in gpgpu computing :(

  • @worldsking3 Furthermore, if you were to be picky you could argue the "lost" 3% is down the fact that this isn't a controlled comparison - the user is changing the scene conditions with the mouse so it won't be a 100% fair comparison between the two scaling setups...

  • @worldsking3 That's 96.7% scaling efficiency - I would say that's as proportional as your ever going to get.

  • sadly amd decided to beat a dead horse with a stick (x86). WHY AMD WHY! CSIC does not scale for NUMA environments!

  • My question is, where can I download the OpenCL enabled Powder Toy?

  • It didn't scale proportionately from 18 cores to 24 cores. At 18, it was 5.26 ms and at 24 its 4.08 ms. Moral is to stop at 18 cores!!!

View all Comments »
Loading...
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