core speeds vs actual work accomplished

[polluxmr2]

I have a question. ( and its somewhat n00bish). how has the ability of procs to do complex mathematics and such increased over the past several years when their speed in GHz really hasn't gone up all that much.

for instance, my 1.6 ghz intel ( some cheap POS ) cpu can do a "discover dengue drugs" task ( on one core ) in about 1.3 hours. my 1.8 ghz 2500+ barton core ( o/c'd to 2.2 ghz ) takes just over two hours. I understand there some architecture differences, but there has to be something more at play? instruction sets? FSB? what?

[NeoGen]

There's so many factors that add up for the final performance... I can think of CPU speed, FSB speed, RAM speed, RAM latency, cache size, L2 cache size, L2 cache speed, special instructions (MMX/SSE/SSE2/etc...), cpu pre-fetching efficiency, number of cpu execution pipelines, number of FPUs within the cpu.... and I'm sure there might be even more that I'm not remembering right now... (And I'm not entering the realm of the I/O devices)

So you see, there has to be a balance between all these factors for the whole to work faster. For example, improving memory speed will only help you until a certain point, if the Front Side Bus can't keep up with the memory output, the memory will just have to work at a slower pace no matter how fast it is in the specs.