P4 2.4C OC questions - very basic.

bdunosk

Senior member
Sep 26, 2000
573
2
81
I'm a bit confused, hopefully someone here will know about how the RAM is involved in overclocking the 2.4C. I understand that the multiplier is fixed, and that all overclocking is done via the FSB. However, are there adjustable ratios available for overclocking the FSB? Is the RAM stuck as a multiplier of 4 of the FSB? Such as...

250 x 4 = 1000MHz FSB (RAM must be 250?)

And if there are adjustable ratios, I'm assuming it will decrease the performance because the RAM will be asynchronous from the FSB?
 

anomaly

Senior member
Nov 14, 2002
401
0
0
P4's can run asynchronyous without a huge performance loss. The main thing is having the highest FSB for the cpu possible, and the highest FSB for the ram. If you can do 285-290 w/ the cpu (which some people have been able to) and 232 with the ram, do it.
 

NoVo

Senior member
May 16, 2001
463
0
0
If you are running an I865 (Springdale) or I875 (Canterwood) board with your 800FSB processor, you will have ratio choices.

If you run synchronously (1:1), your processor and ram will scale synchronously.
ex: FSB set to 250 would yield (2.4c processor)"
CPU=3Ghz -- Memory=500Mhz

Since it's pretty hard to get DDR to run at 500Mhz (especially without realxing the timings), a better option would be setting mem/cpu ratio to 5:4 (aynchronously).
ex: 5:4 ratio (2.4c processor):
250-CPU and 200-MEM or
CPU=3Ghz -- Memory=400Mhz

Hope this example helps.
 

bdunosk

Senior member
Sep 26, 2000
573
2
81
Yes, it does... thanks. So it's not a huge performance penalty by doing this? Is there any data on performance by running asynchronous FSB:RAM?

Now if the prices would only drop a bit, I would consider going to Intel again. :p
 

NoVo

Senior member
May 16, 2001
463
0
0
Originally posted by: bdunosk
And if there are adjustable ratios, I'm assuming it will decrease the performance because the RAM will be asynchronous from the FSB?

There is an in-depth test of the 865 and 875 boards in a forum that I would rather not link to, that suggests a very minimal performance decrease between synchronous and asynchronous ratios. Somewhere around a 2% average decrease going from 1:1 to 5:4.


 

gramboh

Platinum Member
May 3, 2003
2,207
0
0
Ah I just asked about this in another thread, whoops. Any chance you could PM the link? I'm quite curious as to exactly how much performance is lost with async on an 875 platform. Trying to decide between a Barton 2500+ or a P4 2.4C, either paired with TwinX 3200LL memory. I know the Intel platform is pretty much superior once you get to 3GHz+ (which the 2.4 should do) but the CPU+mainboard cost ~$130+ more here in Canada for some dumb reason.