In the 80's and early 90's all the x86 processor equipt computers started using math co-processors and when the Pentium's were introduced they simply disappeared, why? Now Itanium is almost here and we see co-processors introduced again. Why don't we stop looking at higher frequencies and start paying attention to lower level processing functions, so the CPU is not overwhelmed with all of the processing it never had to do after the 486?

The math co-processor was integrated into the 486 DX and above CPU chip. As for why they are separate with the Itanium, I wasn't aware of that fact.

This kind of sound like display adapters and 3D accelerators. At first they were 2 seperate things, but now they combined and faster. But the fact is that there are muany more instructions and a lot more transistors. And Pentium and AMD processors do have the 3DNow! and SSE instructions and transistors on the chip, which is kindof like a co-processor for 3D stuff.

yea 486DX and above always have a math chip built in, As for the Itanium thing I wasn't aware of an external co-proccesor but it makes sence as you can only sqeeze so many transistors onto a waffer. look at the PIII slot chip with the level 2 cache on two extra chips, later they stuck it back on a single die. bet a few years from now the Itanium IIXP/4Ghz will have it integrated again.