• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Advance OC settings, i7 920 D0

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Norcalsteve

Member
Joined
Sep 19, 2009
Location
Baker, FL
So, i am steady at 4.0ghz 12hrs prime stable (will go for 24, just needed to get work done), and have all my base voltages locked in off of Auto... even the Advanced ones. I have read up on Clock Skews (Normal for IOH and 200ps for CPU) and i have read up a little on CPU Differential Amplitude... I understand that it helps the CPU signal "yell" louder above the natural increase of noise at the higher clocking frequency's to help stabilize signal so nothing gets lost in the noise. But is there such a thing as to high of a CPU Diff Amp? I locked mine in at 800mV, but if i set it to 900-1000mV would it kill anything? cant seem to find an answer.

I only ask, cause once i raise my Bclk 9mhz from 191 to 200 (for 4.2 @ 200*21) i can post into windows and find some what a stability loss even with a lot of added voltage.... 1.31v for 4.0 and a bump to 1.35v for 4.2, (it could also be my QPI, that is at 1.27500 which is fine for 4.0).

long story shory, who plays with CPU Diff Amp? and really knows it inside and out?

off topic but while i am hear, i know a BSOD 101 is Vcore and 124 is QPI/Vtt, but what about the weird 7F, A, 9C, and all those weird 1-2 letter/number combo BSOD's, never experienced those with my C0 step chip, just this D0.
 

MIAHALLEN

Senior Member
Joined
Aug 29, 2007
Your understanding of differential amplitude ("DA") is correct. I have played with it a lot, but have never nailed down any measurable increases past 900mV. Usually I start at 800mV, move up to 900mV and then 1000mV CPU DA during benching sessions. And it rarely makes much difference :rolleyes:

On a positive note, there is no chance if this setting causing any harm, so play away and report back if you discover anything useful :)

1.35V for 4.2GHz is excellent, my DO needs 1.425V for that speed :cry:

I don't have an answer for you on your BSOD questions :shrug:
 
OP
Norcalsteve

Norcalsteve

Member
Joined
Sep 19, 2009
Location
Baker, FL
k, i will test out DA at a higher amp when i mess with 4.2ghz more.

As for 4.2 @ 1.35, i am sure its probably more than that, i really need to sit down and test stability, but when i can run Cinebench 4D for 3 passes and test P95 normal and small FFT's for a few passes, then i am in the area of the voltages i need. I am sure if i sit down and really tweak it, i would be more than 1.35, realistically. it could also be my QPI Dram (Vtt) volts too... cause i can post in at 1.32-1.33 its just really unstable (thats when i get those weird BSODs) i just feel the any jump past 1.35 on vcore when i am running 1.31 for 4.0, is not really worth it, i only gain a real small performance increace for only 200mhz with a whole lotta extra voltage. I need this thing for 24/7 use, cause i am considering [email protected]! SCIENCE! YA! :-D
 

xtkxhom3r

Member
Joined
May 29, 2008
Location
San Antonio,TX
hey would you mind posting some of your voltages and advanced settings? iv always wondered what to put for some of those advanced settings like the DA and the clock sqews...where did you read up on the da and sqews?
 
OP
Norcalsteve

Norcalsteve

Member
Joined
Sep 19, 2009
Location
Baker, FL
(about 12 hrs into P95 going for 24, so the settings below are from memory, I will reboot tonight to verify and edit if necessary)

Normal voltages
4.0 w/HT @ 1.31 vcore
QPI Dram (VTT)- 1.27v
IOH's - 1.20/1.50v
ICH's - 1.20/1.50v
PLL - 1.88v
DRAM - 1.64v @ ~1500Mhz; manual timings set to 8-9-9-24-1T. Crap timings, i know... should get Dominators's or OCZ's

Advanced settings
LLC - Enabled
CPU Clock skew - 200ps
CPU Diff Amp - 800mV
IOH Clock skew - Normal

Normal stuff disabled i.e. - PCIE Spread spec, Speed Step, C1E...

If you want screenies of the bios, i gotta find batteries for the camera.
 
OP
Norcalsteve

Norcalsteve

Member
Joined
Sep 19, 2009
Location
Baker, FL
He's cheating

I know it

:D

You caught me!!! i dunno how accurate that is though, like i said, i still need to test stability for 4.2 thoroughly, so if its only my QPI Dram (Vtt) holding me back, then yes, 1.35 it is!! I only upped it from 1.275 to 1.29 (QPI) on the 4.2 trial and got "stable" but didnt test P95 stability. I will mess with QPI at 1.3+ and 1.35 on vcore for that 4.2, then if all is stable, GREAT!

As of now, i got 6 more hours on P95 for my official 24hr 4.0 stability test! ++crosses fingers++ then i will post those settings/screenies for you.
 
OP
Norcalsteve

Norcalsteve

Member
Joined
Sep 19, 2009
Location
Baker, FL
Heres the screenies of my settings. Yeah, i know i said i would run P95 for 24 hrs, but i got antsy, and i want to fiddle with 4.2+ tonight following Miahallen's technique, and waiting 24hrs would be midnight... yeah, its late for me :eek:

===============
 

Attachments

  • 4000mhz stab test.jpg
    4000mhz stab test.jpg
    213.1 KB · Views: 1,811
  • Voltages.JPG
    Voltages.JPG
    184.6 KB · Views: 1,826
  • DRAM Timings.JPG
    DRAM Timings.JPG
    179.7 KB · Views: 1,819
  • Advanced settings.JPG
    Advanced settings.JPG
    188.3 KB · Views: 1,921
OP
Norcalsteve

Norcalsteve

Member
Joined
Sep 19, 2009
Location
Baker, FL
My D0 only needs 1.345v at 4.2GHz.:cool:

Nice! I may only need that too... i need to use Miahallen's instructions to really check my OC voltage settings... atleast that is what i will do for 4.2... hope to get that Vcore down for a happier, cooler CPU!:cool:
 

Elfie

Member
Joined
Nov 20, 2009
I'm running 1.29375v at 4.2 but need 1.53125v for the QPI/VTT to run my RAM at 2000MHz with HT on. With HT off I can drop the vcore to 1.26875v.

I found that a CPU clock skew of 200ps and differential amplitude of 900ma helped reduce the vcore.
 
OP
Norcalsteve

Norcalsteve

Member
Joined
Sep 19, 2009
Location
Baker, FL
wow! you got a gem of a chip! you know, i think i will try opposite for my 4.2 settings... drop my Vcore down, and only up my QPI/Vtt. i haven't messed with anything over 1.35 on the QPI, that could be my problem too... trial and error, what a fun hobby!

Hell, now that i think of it, if i try a high qpi for my 4.0 now, I may just be able to lower my Vcore... try 1.35 instead of 1.27 like i got now. Thank god for ASUS's OC profile's, i can save what i got now for a sure thing in stability, and revert if i change to many settings...
 

nzaneb

Senior Member
Joined
Nov 3, 2008
There are always hidden gems. Do you know the batch number on your chip? I've got one that will do 4.2ghz 21x200 bclk on 1.275 vCore 1.25 VTT... (That is with Vdroop disabled though) Have you tried to see how high she'll go yet?
 

Elfie

Member
Joined
Nov 20, 2009
Batch No. is 3939B278

Highest I've been so far is just 4.44GHz with 1.37v and HT off. However since I later found the clock skew and amplitude settings that seem to work I haven't tried higher again just yet. I'll probably try it again over the weekend.
 

kusakabe

New Member
Joined
Apr 8, 2010
old ish thread lol, but my i7 runs 4.2ghz stable with 1.35vcore, 1.96pll, 1.55qpi, and 1.66dram. all with ht enabled. it seems to me that the faster the ram the higher the qpi volt you need, my ram runs at 2005mhz and i can't get into windows with a lower qpi

i have gotten my i7 up to 4.599ghz with 219x21 with my vcore around the 1.47-1.55 range

i plan to try your advance settings to see if i can get it higher though
 

drdeathx

Disabled
Joined
Mar 9, 2009
Location
Chicago Area
old ish thread lol, but my i7 runs 4.2ghz stable with 1.35vcore, 1.96pll, 1.55qpi, and 1.66dram. all with ht enabled. it seems to me that the faster the ram the higher the qpi volt you need, my ram runs at 2005mhz and i can't get into windows with a lower qpi

i have gotten my i7 up to 4.599ghz with 219x21 with my vcore around the 1.47-1.55 range

i plan to try your advance settings to see if i can get it higher though

High overclocks(not stable) and voltages really mean nothing to compare with accomplishing a stable overclock. I got mine to 4.62GHz with 1.42V...........
 

MIAHALLEN

Senior Member
Joined
Aug 29, 2007
High overclocks(not stable) and voltages really mean nothing to compare with accomplishing a stable overclock. I got mine to 4.62GHz with 1.42V...........

It's all in your perspective...High "stable" OC's mean nothing to me compared to High OC's "not stable"...

3D0303.jpg

....but each to his own.

I don't go around raining on others parades just because I have different priorities :rolleyes:
 

drdeathx

Disabled
Joined
Mar 9, 2009
Location
Chicago Area
It's all in your perspective...High "stable" OC's mean nothing to me compared to High OC's "not stable"...

3D0303.jpg

....but each to his own.

I don't go around raining on others parades just because I have different priorities :rolleyes:

Funny. This guy is obviously working on a stable OC. The other guy threw a suicide in the thread so I just made a simple comment that suicides really are worthless. My intention was not to rain on anyones parade as you see...Now we are going into a thread hi jack????? BTW nice High overclock! Too bad it could not be used 24/7! Ahhh its a 980.. Isn't this a 920 thread?