fbs better then higher mult?

[PimpDaddy]

my TB 800 is at 6.5x135.
witch is like 878. I had it on my older mobo at 969 (9.5x102) witch is better should i drop the 6.5 down lower and add more fbs or less fbs and more mult?

[DaveB]

Yes, absoutely! Try to keep the FSB as high as you can for maximum system performance. The multiplier has nothing to do with performance at all, except that when you increase it, you must decrease the FSB which isn't good.

If the CPU runs at 969 MHz, why not try 7 x 133 = 933 MHz or 7 x 138 = 966 MHz?

[dgk]

I think I know enough to voice an opinion on this thread.

I don't agree that the multiplier has nothing to do with system performance. The multiplier controls the speed of the CPU relative to the bus speed. A CPU that runs 10 times the bus speed is processing data much faster than a CPU running at 8 times the same bus speed.

For CPU intensive tasks, this would seem more important that increasing the FSB. I would think that graphic programs, SETI, FoldingAtHome, and such would benefit most.

Raising the FSB, on the other hand, jacks up the CPU speed also, because the multiplier is now multiplying based on a faster bus speed. However, all of the peripherals are also running faster, so tasks that require peripherals more than CPU activity benefit most from this. Maybe database apps would benefit here.

I can't see why disk drives and such don't explode when the FSB goes so high. That sounds like another thread is needed. I'll start it now.

[proze]

the fsb speed is how fast your proc 'communicates' with memory. this is VITAL to processor speed, cos it can only perform if it has the data it needs. a higher fsb is preferable to a higher multiplier any day.. okay, within reason that is... use your common sense.. dropping the mult by 2x to get 2mhz extra on your fsb is silly. but i think you get the idea.

the reason why hdd and what-not don't blow-up is cos pci devices work off a fraction of the fsb. some mobos can set that divider, and some do it automatically. for instance, on most of the newer kt133a boards, as soon as you go over 120mhz fsb, the board starts using the 1/4 divider instead of the 1/3 divider. that means that at 133mhz the pci bus is 33mhz, which is standard. i.e. no blowing up will be happening. :)

[wild_andy_c]

More FSB, less multiplier if you can - all the way. I've conducted some very conclusive benchmarking upto 150Mhz from 100Mhz for a web article. Most interesting.

[Yomama]

dgk (Mar 06, 2001 12:29 p.m.):
I think I know enough to voice an opinion on this thread.

I don't agree that the multiplier has nothing to do with system performance. The multiplier controls the speed of the CPU relative to the bus speed. A CPU that runs 10 times the bus speed is processing data much faster than a CPU running at 8 times the same bus speed.

...


Actually, at the same CPU speed (multiplier x FSB) a lower multiplier and a higher FSB is preferable to a higher multiplier and a lower FSB for two reasons:

1 - with higher FSB, the the memory throughput rises, which moves data to and from the CPU faster.

2 - this one is less obvious, but important too. It is latency: A small chunk of data may have been processed after one cycle, but it still has to wait for another N cycles until it can leave onto the memory bus. The lower N is, the sooner it will be able to get onto the memory bus.

Computing intensive applications like SETI ultimately benefit tremendously from a lower multiplier at the same speed.

Yo

[Vector]

Yo, kind of cpu (stock I mean) are you running? 1.0, 1.1, 1.2....etc

I'm running a 1.1 T-bird @ 1269 (141x9) but don't know if I want to hold it at above 1300 for any length of time because i'm concerned about heat...
Probably shouldn't be though, as my cpu is happily chugging along at 1.67v without a glitch :)

Also don't know how it will react when I close the case...

[Vector]

Yo, kind of cpu (stock I mean) are you running? 1.0, 1.1, 1.2....etc

I'm running a 1.1 T-bird @ 1269 (141x9) but don't know if I want to hold it at above 1300 for any length of time because i'm concerned about heat...
Probably shouldn't be though, as my cpu is happily chugging along at 1.67v without a glitch :)

Also don't know how it will react when I close the case...