anyone running 5GHz 24/7 on their SB system?

jak731 · 27 Mar 2011 at 13:52

1Day said:
There is a whole heap of people who have 5GHz + over-clocks LINX stable it seems.

http://www.xtremesystems.org/forums/showthread.php?t=239175

Simply amazing what they have managed

How come cpus running at the same frequency have different performance? Especially as SB is primarily oced with the multi. i.e. in that list a 5GHz cpu does over 100Gflops while a 5.34GHz manages less.

I notice the same even with my E8400 dual core, one core consistently does superpi half a second faster than the other.

The Asgard · 27 Mar 2011 at 17:45

beh* said:
I have also found the relationship between VTT and VCORE is very important. Just a couple of notches can make the differences between stability and failure.

Is this generally or only at 4.8+?

I would have thought this would only effect memory speed.

As always I'm looking for low voltage overclocks so anything that will allow me reduce vcore is interesting.

1kiwi · 27 Mar 2011 at 18:10

The Asgard said:
Is this generally or only at 4.8+?

I would have thought this would only effect memory speed.

As always I'm looking for low voltage overclocks so anything that will allow me reduce vcore is interesting.

Well I found at +4.5Ghz required a bump in VTT. However at 1.43vcore with 1.2VTT blend will crash in minutes. Lowering it to 1.16 gave me 10 hours prime stability. I assume every CPU has it's sweet spot.

I guess what I really wanted to point out however is that IBT and Lynx (which is what I preferred) did not seem to uncover stability problems in the same way that prime did.

I'm not really a fan of running 1.43vcore but the UD4 suffers from bad vdrop and the LLC options at limited with the beta bios. Hopefully I can get it down to 1.4vcore 24/7, perhaps using an offset voltage.

John24 · 28 Mar 2011 at 01:17

beh* said:
I guess what I really wanted to point out however is that IBT and Lynx (which is what I preferred) did not seem to uncover stability problems in the same way that prime did.

I hear you need to use the latest linpacks for SB, maybe the older ones are the reason for that.

Liquid · 28 Mar 2011 at 02:29

Yoga said:
Likewise, if you have a spare 2600 that runs at 5Ghz under 1.4v, drop me a line

How much are you offering lol

mine does 5ghz under 1.4

The Asgard · 28 Mar 2011 at 09:50

beh* said:
Well I found at +4.5Ghz required a bump in VTT. However at 1.43vcore with 1.2VTT blend will crash in minutes. Lowering it to 1.16 gave me 10 hours prime stability. I assume every CPU has it's sweet spot.

I guess what I really wanted to point out however is that IBT and Lynx (which is what I preferred) did not seem to uncover stability problems in the same way that prime did.

I'm not really a fan of running 1.43vcore but the UD4 suffers from bad vdrop and the LLC options at limited with the beta bios. Hopefully I can get it down to 1.4vcore 24/7, perhaps using an offset voltage.

Had a quick play with VTT last night and it didn't seem to have any effect on stability with higher clocks on lower voltages. Saying that my stock volts of 0.96 seem lower than yours. I certainly didn't take it to 1.2.

I'm currently running 4.5 with 1.29 volts and the chip seems to scale well with volts. Looks like I have an average 26K. Certainly very pleased with performance.