DirectX 11 Compute Shader Leads to Speed Up, Sharing is a "Killer Feature"

You can’t fault AMD’s Richard Huddy for being a little enthusiastic about Microsoft’s upcoming DirectX 11 and the features it brings along for graphics cards. The graphics API’s "Local Data Share" capability is, in his words, the killer feature which should allow for some impressive performance improvement.

Different program threads running on the card will be able to access information from each other quickly compared to past implementations leading to "up to three times the performance possible with DirectX 10.1"

Speedy! Although it’d be nice to know what performance implications this means for GPGPU performance.

Source: PC Games Hardware

• Bookmark on Delicious
• Digg this post
• Recommend on Facebook
• Buzz it up
• Share on Linkedin
• share via Reddit
• Share with Stumblers
• Tweet about it
• Tell a friend

Tagged as: AMD, Compute Shader, DirectX 11, GPGPU, Local Data Share, Microsoft Windows 7, Microsoft Windows Vista, Radeon HD

by Rafael Hernandez on July 15, 2009 · 0 comments