AnsweredAssumed Answered

Using the Target Independent Rasterization (TIR) hardware outside of Direct2D (e.g. from OpenGL or Mantle)

Question asked by respawned on Aug 10, 2014

So AMD's GNC-based chips unlike Nvidia (with Kepler anyway) supports TIR. (Background on TIR can be found in a blog post of the Windows 8 team and in more details are in a Microsoft patent application.) Of course I can use TIR from Direct2D since AMD basically made the feature at Microsoft's request (see the aforementioned blog); the Direct2D API for using TIR is on this MSDN page.


But what about using the TIR hardware on Linux or on Windows but from a different graphics API, for example OpenGL or Mantle? Does AMD support this, and if so where can I find the API specific calls to use?


I should note here for comparison that Nvidia has their own OpenGL method for accelerating 2D vector graphics, called NV_path_rendering. The implementation of NV_path_rendering is substantially different from how Direct2D operates. For anyone interseted in the details, the Siggraph 2012 paper by Kilgard and Bolz explains the differences.


(By the way, AMD should fix the JavaScript on this site. I locks up Google Chrome at 100% CPU utilization after a few minutes, probably when it tries to auto-save, and thus makes it impossible to write a longer post. I had to retype this in a standalone text editor and then paste it here before the site's Javascript-based editor could lock up again.)


Message was edited by: Respawned Fluff: added the patent app link