| bdbdbd said: |
CISC (Complex Instruction Set Computers) generally are CPUs that include instructions which do multiple instructions at the same time in order to increase processing power; much like my theoritical matrix operations earlier. Increasing the number of instructions (and adding complex instructions) has become less popular today mainly because in only adds processing power to specific tasks rather than adding processing power in general. Traditionally, when designing a processor for a specific application (like a game console) the favourite approach has been to add complicated instruction sets because there are usually only a few areas which need dramatic improvments.
RISC (Reduced Instruction Set Computers) generally are CPUs that have a very small instruction set and gain performance through increasing the clock speed; in extreme cases (super scaler RISC processors) the instructions can be reduced to integer bit-shifts, addition and memory moves. RISC processors were usually found to be much faster at running code in general and (in many cases) clever programmers could use tricks to outperform CISC processors' built in instructions.
The obvious example of the advantage CISC processors have over RISC processors is the GPU in most computers. Your GPU can perform far (FAR) more Floating point operations per second than your CPU (meaning it can produce far better graphics than your CPU could) but you would have far (FAR) worse performance if you attempted to run Microsoft Word on your GPU then your CPU.







