On Mon, Jun 07, 2004 at 03:09:15PM -0400, Sean Estabrooks wrote: > "T. 'Nifty New Hat' Mitchell" <mitch48@xxxxxxxxxxxxx> wrote: > > > Consider the mmx class of instructions that Intel introduced. > > What "C" code causes these instructions to be generated by gcc. > > > > See ..../arch/i386/lib/mmx.c .... > > One of the problems with complex instruction set processors (GFX > > engines are complex) is that compiler writers often ignore the full > > instruction set.... ... > > > > These are not frame buffers with 'common' processors glued to them. > > > > What are you talking about? Do you know how many open > source 3D drivers there are? Please check out: > > http://dri.sourceforge.net/ Good pointer. Looking at the design documents I see mention of what I am saying. These are not simple frame buffers with common processors glued to them. The design of dri shows awareness of this task and associated issues. What I am failing to communicate is that there is a fun set of issues associated with the use of asymmetric special purpose co-processor subsystems for acceleration of any type. Of interest dri does work on FC2 (nv driver) with my nVidia Corporation NV11 [GeForce2 MX/MX 400] card, glxgears and more. Just not as fast as the accelerated driver. -- T o m M i t c h e l l /dev/null the ultimate in secure storage.
