GB Z77X-UP7 - lower RAM overclocking

[Blue Storm]

Sorry for double post...

 

 

XMP working fine now but found something weird while setting XMP...

 

 

Really 1.4V VTT? I rather set 1.15V because I think that's very high voltage for air don't want to kill my Ivy:confused:

[Massman]

I think that's Corsair's fault for incorrectly configuring the XMP profile? Just loaded up F5f here with GSKill RipjawZ and 'only' get 1.25V VTT by XMP.

 

[Blue Storm]

Hmm yep it looks like Corsair's fault...

[Massman]

Ah, just did some measurement - the 1.25V or 1.40V is just read-out not actually setting that voltage. If I load the XMP here (1.25V) it will just be default (1.05V) real.

 

Nothing to worry about, in other words

[Blue Storm]

I didn't have courage to try it:Dthanks for information;)

[Leeghoofd]

what about 16GB Blue Storm does that work now too ?

 

FYI same weird VTT detected on Asus and MSI boards...

Edited November 1, 2012 by Leeghoofd

[Blue Storm]

yep works too;)

[stasio]

Ah, just did some measurement - the 1.25V or 1.40V is just read-out not actually setting that voltage. If I load the XMP here (1.25V) it will just be default (1.05V) real.

 

Nothing to worry about, in other words

 

Yea,as I noted in my previous post,Vtt is always 1.05V.

XMP set Vdimm 1.65 and Vtt 1.35,but in Windows Vdimm is 1.680V (UP5).

 

 

Edited November 1, 2012 by stasio

[Guest Bullant]

Hey Blue ,

Have you tried PSC on new bios to see if improvement?

[Blue Storm]

I didn't try PSC on UP7 earlier so I don't have result for compare...but I will try it to see how good it is

[Blue Storm]

It looks that it still has problem with PSC...when I set 2600 or 2666 multiplier with 11-12-11 2T which was stable on UD5H or P8Z77-V Pro I have 51 on post display...the settings are the same

 

//By the way...XMP with Dominators Platinum 2400 CL9 working also very good

[Guest Bullant]

It looks that it still has problem with PSC...when I set 2600 or 2666 multiplier with 11-12-11 2T which was stable on UD5H or P8Z77-V Pro I have 51 on post display...the settings are the same

 

//By the way...XMP with Dominators Platinum 2400 CL9 working also very good

 

Thanks for taking the time to check mate.

[M.Beier]

Tested 6-layer board and went well over DDR3-3000. More layers = more complexity = need finer tuning. That and some other stuff.

 

OH great!

Banana is better than chocolate, because you prefered bananasplit next to chocolatemouse?

Pieter, sigh, seriously that is one dumb conclusion you've reached.

- You do know there is a lot more parameters, dont make such a bold statement.

 

Furthermore no, more layers is NOT same as more complex, in fact it is much more simple as less other components or traces block the path, hence; you can cross through layers.

[sin0822]

OH great!

Banana is better than chocolate, because you prefered bananasplit next to chocolatemouse?

Pieter, sigh, seriously that is one dumb conclusion you've reached.

- You do know there is a lot more parameters, dont make such a bold statement.

 

Furthermore no, more layers is NOT same as more complex, in fact it is much more simple as less other components or traces block the path, hence; you can cross through layers.

 

I feel like pulling a Tin on you.

[ME4ME]

for our amusement, please do

[zeneffect]

more is always more complex... thus the meaning of the word more.

[I.nfraR.ed]

Might be a stupid question, but how to enable vcore higher than 1.85V on UP7?

I'm running F4 bios, single bios switch enabled. I don't see any option in bios to enable extreme voltage or some sort of jumper.

[Massman]

OH great!

Banana is better than chocolate, because you prefered bananasplit next to chocolatemouse?

Pieter, sigh, seriously that is one dumb conclusion you've reached.

- You do know there is a lot more parameters, dont make such a bold statement.

 

Furthermore no, more layers is NOT same as more complex, in fact it is much more simple as less other components or traces block the path, hence; you can cross through layers.

 

Sigh. Yeah ...

[pro]

I feel like pulling a Tin on you.

 

[sin0822]

Here is a new BIOS with much improvement:

http://www.mediafire.com/?st532240qeslhwi

BIOS F5G

It adds read and write slew

It helps CL12 compatibility

 

Here are the new settings and what seems to be the best for different types of modules:

 

So start with level 2 and then increase for different levels of stability and such for samsung and PCS, samsung will for instance give you instance 51 on level 1 if you go too higher over XMP multiplier, but change to level 2 and it will not 51 and it will boot.

[Massman]

Just to clarify: F5g BIOS is not the one for improved PSC/BBSE overclocking - it's just the one fixing the tCL=12 incompatibility issues. The Slew Rate finetuning is mainly to help with bootup/post procedure but won't add additional overclocking headroom.

 

Afaik PCB/BBSE bios is coming later.

 

F5f is working out well for memory overclocking here:

 

[l0ud_sil3nc3]

What is the BBSE problem you speak of, this board runs BBSE without a hitch no need for fixing that as PSC is what is really broken.

 

Did not test MFR/CFR/Samsung so can't comment on that but just need PSC 2600+ to work and work well and that is all that is lacking imho.

[sin0822]

the slew rates give me higher headroom on samsung and PSC from my findings, F5F can give you the same headroom on samsung IMO.

but yea this BIOS isn't mean to help PSC or BBSE, however the increased slew rated helps a lot with PSC, gave me an extra multiplier to work with, but it does help boot with higher multipliers which is part of the battle.

F5F detects my samsung modules and changes its slew rate on the fly however the slew rates are very helpful for both samsung and PCS, and I think they should be an option for users to change, as it can benefit and what if the board doesn't detect it?

Edited November 12, 2012 by sin0822

[Massman]

What is the BBSE problem you speak of, this board runs BBSE without a hitch no need for fixing that as PSC is what is really broken.

 

Did not test MFR/CFR/Samsung so can't comment on that but just need PSC 2600+ to work and work well and that is all that is lacking imho.

 

Oh - BBSE is working properly? Like in DDR3-2600+ C8-12 on air cooling? It's been a while since I tested BBSE memory on the UP7, so I might have missed the improvements made .

 

Good to hear!

[zeneffect]

BBSE has been working since at least F3.

 

F5J seems to work pretty well with PSC. was giving some crucial ballistix a run @ 2600 8-12 and it seems to be pretty solid for the most part. only issue was getting 2600 multi to work, would always 51 no matter what, however I can 2400 multi then bclk up with no issue now

 

*edit*

 

switched to sammy... bios settings wont take unless cmos is cleared and set on the initial boot. set c11, still stuck @ c10. i need to re-flash and test this again because i think it works fine with psc and may be a samsung issue.

 

*edit edit*

 

reproduced issue with PSC, reflashed, issue went away.

Edited November 23, 2012 by zeneffect