wanted to know how many times 386 is weaker than present pc
start with clock speed, buss speed, memory speed, and something like number of clock cycles for an add , mul, or floating point div. fetch times from memory might (or indirect addressing clock times) might be useful too.
then you have to figure in multi-core and cacheing on the new pc. good luck with that one. if you know how the chip and cache work, you could probably pencil and paper figure out some test scenarios and get some numbers that way (IE by theoretical derivation).
and then there's the whole issue of expanded instruction sets on the newer processor. it may be able to do things more efficiently whit newer instructions such as mmx, simd, etc. this is where exactly what you want to measure comes into play. so you'll need to measure a number of things to get a general idea of how much faster the new pc is at different things. there may even be a difference in the number of instructions (as well as their clock times) required for something as simple as a fetch, add, and store.
unless its mission critical, you might just want to say the answer is "its boatloads faster!" and get back to building games. <g>
if you have access to both pc's, simply decide exactly what you want to measure, exactly how to measure it on both pc's (this will probably be a compromise at best), and write a little benchmark. that will take into account all the new stuff on the newer pc, except perhaps for using special instruction sets. and if what you want to measure can use new instruction sets, then you can simply add a test using the new instruction set to your benchmark, which will give you an idea of the performance of the new vs old instruction set..
