View Full Version : Northbridge Core Clock, FACT or FICTION?
CoW]8(0)
10-06-2007, 02:57 PM
Does this mysterious equation for calculating your NBCC hold true?
NBCC = FSB * (Default CPU Multiplier)/(Set CPU Multiplier)
Or is it a lie spread by those will bad OC'ing boards?
CoW]8(0)
10-06-2007, 03:05 PM
A thought experiment (I'm not near my machine right now, so someone else could give it a shot):
1)With the stock multiplier of your CPU, find the maximum OC you can achieve.
2)Then decrease your multiplier by 1
3)Keep decreasing until you fail to boot at those settings
In theory, this will only be raising your NBCC according to the equation above.
For example, say the maximum OC of your E6600 was 3600Mhz (@ 9x multiplier, 400Mhz FSB).
If you went all the way down to a 6x multiplier, you would be running your NBCC at (9/6)(400) = 600Mhz. You should crash at 6x since there aren't any boards I know of that have a NB that hits 600Mhz.
No need to go through all that. Sandra shows the real speed.
http://www.thetechrepository.com/showthread.php?t=30
CoW]8(0)
10-06-2007, 03:23 PM
That's what makes me a skeptic, Sandra is the only program that shows the 'true' speed. No other program seems to do this.
And if it's true, then the OC ability of several boards come back into question. The P35, X38, 680i, P965 would all have to be reconsidered for OC ability.
For example, it would seem that my motherboard (680i LT) is a MONSTER overclocker, since it can boot at 400Mhz FSB on the 6x multi of my E4400 (400*10/6 = 666Mhz).
Or should I say, my motherboard is a beast :)
first off, it was only decided that this formula worked for P965 i believe, before 680i was even out so no...your board isn't a monster.
second, unless there were changes in the BIOS by Asus that eliminated this NB overclocking...it never was really true otherwise...
click for full size...all air, stock board, stock cooling on the board, no tricks...
http://img.photobucket.com/albums/v220/3oh6/ocx/629b_e6600/550x6_ultra-120_prime-1sm.png (http://img.photobucket.com/albums/v220/3oh6/ocx/629b_e6600/550x6_ultra-120_prime-1.png)
all three P5B-Dlx i had could run 550*9/6 = 825Mhz. or the NBCC frequency isn't exactly like FSB frequency and the boards did run that high. you have to understand that the article was written when P965 was still fairly new and i think at the time, we didn't even know about a C2D FSB wall that is only extended by PLL voltage or sub-zero cooling.
CoW]8(0)
10-06-2007, 04:44 PM
So was this NBCC formula true for the P965? Because I'm not sure what you're saying.
8(0);2474719']So was this NBCC formula true for the P965? Because I'm not sure what you're saying.
look at the screen shots...E6600, stock multi of 9...priming at 550*6. so according to the equation, FSB*original multi/running multi (550*9/6), my NB is actually running at 825MHz in the screen shot above. i highly doubt that.
so either Asus has changed the BIOS so that changing the multi doesn't overclock the NBCC, or some other factor is involved, or the whole idea of lowering the multi overclocking the NB is not true.
CoW]8(0)
10-06-2007, 05:10 PM
Thanks for clearing that up. I was thinking the whole idea of overclocking the NB when you lower the multi didn't make any sense logically.
That's why I'm leaning towards the notion that Sandra just had a programming mistake and the NBCC = FSB.
And is increasing the PLL voltage compensating for a CPU FSB Wall or a motherboard FSB wall?
Hutch
10-06-2007, 10:40 PM
8(0);2474764']Thanks for clearing that up. I was thinking the whole idea of overclocking the NB when you lower the multi didn't make any sense logically.
That's why I'm leaning towards the notion that Sandra just had a programming mistake and the NBCC = FSB.
And is increasing the PLL voltage compensating for a CPU FSB Wall or a motherboard FSB wall?
The NBCC was dismissed by the person who made the hypothesis of it. Increasing your CPU PLL Voltage will help overcome a CPU fsb wall.
The NBCC was dismissed by the person who made the hypothesis of it.
figured as much...i didn't know he did, my bad. why don't they pull it from the TR i wonder?
CPU-Z screenshots have no bearing on this discussion. FSB and NBCC are not one and the same.
mach82
10-07-2007, 02:59 AM
8(0);2474537']Does this mysterious equation for calculating your NBCC hold true?
It's not true. The NBCC theory has never been proven to be right. SiSoft Sandra has misled many people - it's just the Sandra's bug, nothing more The same bug can be observed in CoreTemp when you lower your CPU's multi.