Any good benchmarking on FSB vs. CPU speed?

[Rio Rebel]

I know this isn't new, but I have been searching off and on for a couple days now and I can't find a good article comparing FSB increases to CPU speed increases. I'd love to get a better feel for how a T-bred (or thunderbird, or Barton, or whatever) at, say, 166 x 12.5 (2075) would compare to 200 x 10 (2000). Just how much faster would a cpu on a 166mhz bus have to be to equal one on a 200mhz bus? Or does it actually scale?

The one half-decent article I found seemed to indicate that games at higher resolutions gain virtually nothing from a higher FSB speed.

Any input?

 

[Boogak]

Originally posted by: Rio Rebel
I know this isn't new, but I have been searching off and on for a couple days now and I can't find a good article comparing FSB increases to CPU speed increases. I'd love to get a better feel for how a T-bred (or thunderbird, or Barton, or whatever) at, say, 166 x 12.5 (2075) would compare to 200 x 10 (2000). Just how much faster would a cpu on a 166mhz bus have to be to equal one on a 200mhz bus? Or does it actually scale?

The one half-decent article I found seemed to indicate that games at higher resolutions gain virtually nothing from a higher FSB speed.

Any input?

This doesn't answer your main question but gaming at high resolutions is usually video card limited rather than CPU limited. That would explain why there's no gain in that article.

 

[Actaeon]

Originally posted by: Rio Rebel
I know this isn't new, but I have been searching off and on for a couple days now and I can't find a good article comparing FSB increases to CPU speed increases. I'd love to get a better feel for how a T-bred (or thunderbird, or Barton, or whatever) at, say, 166 x 12.5 (2075) would compare to 200 x 10 (2000). Just how much faster would a cpu on a 166mhz bus have to be to equal one on a 200mhz bus? Or does it actually scale?

The one half-decent article I found seemed to indicate that games at higher resolutions gain virtually nothing from a higher FSB speed.

Any input?

Not really sure on such reviews.

Though I believe the 200mhz FSB @ 2000 would probally be faster.

EDIT: There is no reason why you shouldn't be able to balance your CPU multiplier (unless you got an Intel processor ), and whatever FSB in order to get that last MHZ out while still obtaining high FSB speeds.

 

[Rio Rebel]

Thanks for the responses.

That makes perfect sense about the higher resolutions being gpu rather than cpu limited. Somehow that slipped past me.

As to Actaeon's last statement, it's generally a good point, but I can't implement it. I'm running 200FSB x 10.5 (2100) on an MSI nForce2 board. I every jump in mulitplier options (.5) means 100mhz at this point. And the FSB options top out at 200, so I can't increase the speed there.

Then there's another catch: I'd prefer to run the memory synchronous with the fsb speed. So if I do leave it synch, every drop in FSB speed also slows down the memory as well.

I suspect I'd be better off dropping to 195 x 11 (2145), and I'm sure this isn't a significant difference either way. I'm more curious than anything else.