Back to Article

  • Jafar - Thursday, March 28, 2013 - link

    I could be wrong, but AMD's zero copy will only see light this year, although it has been announced for a while. Reply
  • TeXWiller - Thursday, March 28, 2013 - link

    Here you go: Reply
  • Jafar - Thursday, March 28, 2013 - link

    Thanks TeX!

    Part of the confusion I had is probably due to vague memory from AFDS12, where AMD announced their HSA initiative, and the unified memory architecture. Zero Copy might mean different things ( ). I guess what AMD is trying to deliver this year and the next is CPU/GPU with "equal" rights to memory access. I believe current APUs still set aside part of the memory to be exclusive to the GPU.
  • iwodo - Thursday, March 28, 2013 - link

    Cant wait to see how haswell GT3 with cache to perform. I am hoping for a surprise. Reply
  • A5 - Thursday, March 28, 2013 - link

    Well, it's probably a big threat to the GT 640Ms of the world (which is great for an iGPU, mind you), but don't expect desktop-gaming level performance. Reply
  • Shadowmaster625 - Thursday, March 28, 2013 - link

    ack! That van looks like it is photoshopped into that image, it looks like it is floating in the air above the fence rather than on the ground. What is this, a game from 1997? Reply
  • Mr Perfect - Thursday, March 28, 2013 - link

    Those who live in glass houses... Reply
  • yefi - Saturday, March 30, 2013 - link

    See everything? Reply
  • xenol - Wednesday, April 03, 2013 - link

    Is it bad that I'm finding it amusing that zero copy is just being implemented on higher performance machines when it's been a staple in embedded programming for years?

    The framework I use handles events through a zero copy mechanism so as to avoid unnecessary memory consumption.

Log in

Don't have an account? Sign up now