A feature on newer intel pentium processors. Most processor functions use one opcode (instruction) to operate on one set of data, so the CPU reads the opcode, loads the data, performs the operation, and outputs the result. lather, rinse, repeat at 200MHz. MMX loads one opcode and performs the same operation on a whole pipeline full of data. Since MMX loads the instruction once instead of every time, it can increase CPU speed for repetative functions like 3D rendering and MP3 encoding or decoding. My quick and dirty test indicates that using MMX opcodes gives about a 10% speed up in MP3 encoding. YMMV

MMX, unofficially "MultiMedia eXtensions" (although Intel won't admit that), is a set of extra opcodes to perform several operations very fast, especially useful in graphics and sound processing. You can do stuff such as adding, subtracting and multiplying 8x8bytes at a time, with or without "saturation" (so that 200 + 100 becomes 255 and doesn't "wrap around").

Of course, you cannot do FPU stuff while you're doing MMX, or you will die horrendously.

According to my notes, MMX is short for Matrix Mathematics eXtension. MMX uses SIMD (Single Instruction Multiple Data) operations, which allegedly gives it an eight-times performance speedup for those things it is ideally suited for. For those who care about such things, the reason MMX is incompatible with FPU operations is because Intel, not wanting to create new registers, decided to map the eight MMX registers to the 64-bit mantissa part of the 80-bit floating point registers.

Log in or register to write something here or to contact authors.