• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Wrong CPU speed / FSB / Vdrop

dolphumous

dolphumous

2[H]4U
Joined
Jul 6, 2004
Messages
2,138
OK. Testing some things out. I'm running 400x7 in the BIOS.
Do the other apps read it @ 3.6GHz because 9x is the default multiplier and they don't know that I lowered it?

I can understand that but why does CoreTemp show a 514.29FSB?
I also noticed I'm getting a little Vdrop. I've got it set to 1.325 in the BIOS.

Temps aren't what I want right now but you can read my ranting in the Ninja/LGA 775 thread a few down from this one.

what.jpg
 
It looks like Orthos and CoreTemp don't know how to deal with lowered CPU multipliers. They detect you have an E6600 and therefore assume the CPU multiplier must be 9. I have seen this myself.
 
Running ram 1:1

One scenario:
400MHz x 7 = 2.8GHz CPU Speed
400MHZ x 8 = 3.2GHz "
400MHz x 9 (default) = 3.6GHz


Or if you really wanted to OC your FSB you'd have to lower your multiplier.

Second Scenario:

550MHz x 9 = 4.95GHz
550MHz x 8 = 4.4GHz
550MHz x 7 = 4.0GHz
550MHz x 6 = 3.6GHz

See how FSB running at a 1 to 1 ratio and the multiplier effect the final CPU speed?
Also, having a higher FSB with the same CPU speed will perform better because the overall system bus is much higher.

(550MHz x 6) > (400MHz x 7)
 
why such a moderate overclock? I'm sure 8x would be just fine with a minimal vcore bump, and 9x400 is likely doable with high end cooling and high vcore.

But anyway, most programs aren't able to see that you're running a lower multiplier (think it's because of the new eist causing multiplier to be applied differently, since nbcc wasn't ever talked about before, and I don't have issues with dropping multipliers with my dothan).

But anyway, orthos is seeing the fsb and default multiplier and applying that to calculate speed. Coretemp must be coded to read the proper multiplier. However, it still pulls clock speed using some function that's not accurate (if you notice in system properties, it's how windows is reading it). So when coretemp goes to break it down, it divides 3600 by 7 and arrives at 514.

Your droop isn't bad at all if you're seeing 1.312 under load when 1.325 is set. Then again, I've noticed droop gets worse with higher vcores.

edit: hmm 1.312 isn't load vcore. Yeah, my new ds3 does that too, but way worse. For example, if I set 1.375v, I get 1.328 idle and 1.280-1.312 under load. With a different ds3 I tried, it seemed to be spot on at idle and droop only a tiny bit when dealing with the lower vcores. With higher vcores, it seems to be just slightly shy of set vcore, but droops terribly. 1.5v set was doing 1.49v idle and 1.44v load.

I did do the droop mod on my rev1 ds3 and it works wonderfully.

edit: http://www.xtremesystems.org/forums/showthread.php?t=110101&highlight=ds3+mod
 
I was thinking about it and everything you said is also what I concluded. Very cool.

It's moderate right now because of the temps and other things I'm learning about.

I said to hell with it when I first got it and it booted @ 400x9 but BSOD'd in about 4 minutes...

Once I figure out how to get 3.6GHz full load under 60C I'll do it.
 
dolphumous said:
I was thinking about it and everything you said is also what I concluded. Very cool.

It's moderate right now because of the temps and other things I'm learning about.

I said to hell with it when I first got it and it booted @ 400x9 but BSOD'd in about 4 minutes...

Once I figure out how to get 3.6GHz full load under 60C I'll do it.
Click to expand...

haha I'm barely doing it with my e6400. I'm running 8x450=3600 with 1.455v load volts and water cooling, and I'm hitting 60'c on the nose under max load.
 
That mod looks interesting.

And my temps...etc I blame the Nijnja so far...

I don't know why they spike so high.
 
hmm ninja isn't too bad. It's bolt mount and not clip right? Maybe ihs is too concave and needs to be lapped. But yeah, running in the 60's with c2d's doesn't scare me at all, assuming coretemps
 
No, it uses the shitty push pins.

I just read somewhere about the IHS can be concave on c2d...wth is up with that?
 
To add to the confusion... Since I finally got EasyTune5 installed that's also giving me different voltages and temperatures.

Which ones should I go by?

Example:
CopeTemp right now says 47 degrees / 1.32V
EasyTune5 says 32 (which is where my old X2 idled in the same case) / 1.36V

Is that using the diode that I'm used to seeing with a A64 system?
 
You must log in or register to reply here.
Back
Top