Le Wed 19 Apr, à 17:21 -0700, James Richard Tyrer a écrit : > Timothy Miller wrote: > >What does it mean to have a MISC stack processor? The MISC idea uses > >special purpose 'register' operands to encode the opcode. Stack > >machines have totally implicit operands. > > The docs say it means: Minimal Instruction Set Computer. I think that > this means what most people think RISC means, a computer that doesn't > have a lot of instructions -- IIUC, that doesn't have a lot of extra > instructions -- more complicated things are achieved by executing > multiple instructions. Since with a FORTH type stack machine (as > opposed to a C type stack machine like HOBBIT) you do not need to > specify the location of the data for operations (just like an HP > calculator the data for the operation is the next N registers on the > stack), except for fetching from the stack, so you achieve the same > result as far as the amount of bits needed for the machine > instructions. You are simply going to have an op code instead of > a register location.
Yes. Some FORTH type stack machine op codes were only 5 bits long. With one 32 bit memory fetch you can load 6 op codes, and they need no operands. This is a natural instruction pipeline. The "hardware" data stack is also a natural data cache. Data stack manipulation instructions are like mov commands, but instead they don't move the lengthy data stored in the registers, they just alter the register order. Optional "modules" like ALU always take their operands form the register labeled "top" and "next" (0 and 1) and put the result back in "top". If the design is such that data stack manipulation instructions can re-order the stack before the ALU gives its result, you acheive the same clock cycle usage as with Timothy Baldridge MISC idea. To me Timothy Baldridge's idea rejects part of the scheduling complexity to the software compiler, but on the other hand, data and instruction cache is still a problem to be resolved in hardware. With Forth-like stack machines, the hard part is the stack manipulation (but really you can do with a few instructions), but you have caches for free. Plus the compiler is straightforward. Robin _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
