TerraRaptor - Rampage Extreme @ 680.1MHz - 680.11 MHz Reference Frequency

[d0minat0r]

Very nice!

[M1RROR]

great! the next step should be cold (Nb)

[Jokot]

Nice, cold nb next i hope :D

[yosarianilives]

Holy shit

[TerraRaptor]

Thank you friends. I will first try to improve on x7 multi superpi result and then will probably put nb on SS. 

Does anybody know if using pci video or single channel/single sided sticks make any difference with fsb?

[GRIFF]

Nice result for air cooling. I'm checked different slot and type of IC but have a same limit. My board at air cooling got 660-665. The problem of this platform is cpu fsb wall. You need good cpu for take 700+

[TerraRaptor]

7 hours ago, GRIFF said:

The problem of this platform is cpu fsb wall.

Well, NB should be the limit here - I was able to reproduce same clocks with e8200 today.  Will try cold later this week.

[Eisbaer798]

Nice Score TerraRaptor! Congrat

[TerraRaptor]

10 hours ago, Eisbaer798 said:

Congrat

Thank you. Tried cold yesterday but no success - I was using the same profile that allowed me to post @675 at positive temparatures on cpu+nb but couldn't even get into windows with 660MHz once I've turned on my SS. Sam is right saying that REX+Wolfdales are not reliable with non-LN2 cooling at high FSB - or there is something we can discover with BarEdit to stabilize it.

[Alpi]

Wow ! Very nice result !

[Eisbaer798]

The Rex boards can be quite tough. With my good board I can boot at positive temperatures with a FSB of 660 without problems. (@ 1.71 volt @ + 27 ° C) Unfortunately, under LN2, the NB does not allow one MHz more. Such a pity, I thought I finally found a 700+ board ;-( Isn't it a bit too risky for you using positive temperatures at 1.8 V for your NB? What do you mean exactly with BarEdit? You can answer via PM if you like.

[TerraRaptor]

2 hours ago, Eisbaer798 said:

Isn't it a bit too risky for you using positive temperatures at 1.8 V for your NB?

Not sure but haven't killed any board with high vNB yet - I guess temperature control is the most essential thing when we talk about vNB as vNB affects temperature directly. Actually, x48 must be very sensitive to temperatures - while searching broken cap I was heating whole board several times and have noticed that ВС resistance of x48 changes dramatically with temperature - it is smth like 8 Ohm @25C, 4 Ohm@60C, <2 Ohm@100C. So, there could exist avalanche effect when NB starts heating rapidly dropping resistance further thus consuming more and more power.

2 hours ago, Eisbaer798 said:

What do you mean exactly with BarEdit?

My guess that random stability issues with high FSB (i.e. same profile is stable on day1 and not stable on day2) may be caused by some hidden settings of x48 - board sets these automatically on powerup. We still can monitor/control  these settings with software called BarEdit (other options may be Pmem, RWEverything etc) . 

Here is a good read on BarEdit>MCHBar settings compiled by Don_Dan - 

I think that if we collect enough dumps of registers from "better-than-usual days" and compare it to "usual days" dumps there is a chance of locating some entries that we can set manually and force stability.  Just an example - DFI boards usually had much wider number of settings because Oscar Wu gave us an ability to manipulate some of those "not in public datasheets" settings.