Timothy Miller wrote:
There are lots of tradeoffs in computer architectures. One of the ways that RISC and VLIW designs get their speed and simplicity is by trading off instruction compactness. VLIW is particularly bad, requiring lots of NOOPs.
A graphics processor is the opposite of RISC, it is (or should be) a VCIC [Very Complex Instruction Computer]. I once worked this out for a general purpose CPU that used code with mnemonics that looked like FORTRAN IV but could be parsed directly into machine instructions.
The theory of this is that it does not use the large instruction stream bandwidth that RISC does and the only branches are those actually in the code. True RISC uses a lot of instructions to accomplish anything and assembler code is based on a lot of jumps and branches. The problem is that the processor runs at a lower clock speed (something that is probably no longer a problem).
I should have patented it then Sun would be paying me royalties for Java. :-D
The actual machine level language (meta-assembler) can then have a higher level language that humans use that can be compiled (without any need for optimization) into the machine level language.
Can we do the same thing with OpenGL? Can we design a low level language that can be directly implemented in hardware that it is possible to compile all of OpenGL into?
-- JRT _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
