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

Dolk's Guide to the Phenom II

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Somewhat. You guys can push for a bit more, but it is true that is a pretty high temp for a low OC, in terms of the 1090T.
 
hi,

i just finished to do some "tests" and here s what i have:

-i have lowered the VCore til the temperature of the cpu never reach 45° under OCCT cpu test (1hour) @3.6ghz (18x20):
1.428v=max 46.5°c
1.418v=max 46.0°C
1.408v=max 45.0°C

-i have set up the multiplier by 0.5pts ....waiting the bsod and watching the temp monitor
test 1= "3dm06" +test 2 = "OCCT cpu test"/12minutes (just to have an idea ;) )

vcore=1.408v (cpu temp @iddle=22/23°c):

3.6ghz(18x20) cpu max temp=45°c
3.7ghz(18.5x20) cpu max temp=45°c
3.8ghz(19x20) cpu max temp=45.5°c
3.9ghz(19.5x20) cpu max temp=46°c
4.0ghz(20x20) cpu max temp=46°c...3DM06 no problems; bsod under OCCT cpu test (+ 6minutes)

i ve backed it to 3.8ghz as my goal is a 7/24 oc @3.8ghz, is it acceptable with thoses temps?:

occt1h38ghz.jpg

thx for help and sorry for the "unconventionnal" oc method...i m a oc newbie :D
 
Last edited:
I'll be getting to your question sometime later. sorry I'm pretty jamed with a couple things atm.
 
just a quickie on voltages with a 1090T

what is the "best" voltage for a 2800MHz NB and a 4GHz CPU?

Currently have NB set at 1.2875v and CPU set at 1.45625v and it is stable, but would like to lower it if poss - what should I aim for or are these good voltages?
 
Seems like a hefty voltage for a 1090T unless you have some pretty deep Vdroop. I've clocked three thubans and all of them required 1.41V under load.

CPU NB seems on par with what I've experienced around that speed.

I'm not the master though. :)
 
Do a 2600 on the CPU-NB unless your memory is at 2000MHz. The difference you see between the two is to low, and in some cases you lose performance.

1.22-1.25v is all you need for the CPU-NB.
 
well after posting i had a fiddle (with the pc not myself) and adjusted the voltages and boosted the cpu multi - left it folding overnight, and still running strong now.

cpu multi now 20.5 = 4.1GHz
voltages now - cpu = 1.4625v, cpu/nb = 1.275v

glad i had a fiddle now!
 
left it folding overnight, and still running strong now.

Not really the most appropriate tool for stability testing. The f@h project depends on accurate results coming back for their work units and an overclocked PC can return inaccurate results. You might find out if your PC locks up through this method, but it won't tell you if the results you are calculating are correct. Much better to run the torture tests on P95 which will tell you that.

Once you know it's P95 stable for 12-24 hours, then you can crank up F@H.

Just my tuppence worth.
 
yep you are right - I thought it was stable but noooo

about half an hour after posting, it crashed, so back to 4GHz...

it was Prime 24hr stable at that speed previously.
 
Seems like a hefty voltage for a 1090T unless you have some pretty deep Vdroop. I've clocked three thubans to 4.0 GHz and all of them required 1.41V under load.
Heh, oops. Left out an important part of my post, see bold & underlined text above

Do a 2600 on the CPU-NB unless your memory is at 2000MHz. The difference you see between the two is to low, and in some cases you lose performance.

1.22-1.25v is all you need for the CPU-NB.
It took a little doing, but I managed to get 2500 on the CPU NB and HT stable with DDR3-2000 memory. FWIW, on a Crosshair IV Formula with the latest beta BIOS, the settings are:

Bus Speed: 250 MHz
HT Link: 2500 MHz
HT Volts: 1.3375 V
CPU-NB Frequency: 2500 MHz
CPU-NB Voltage: 1.26'ish V (with CPU-NB LLC on full)
RAM Frequency: 1000 MHz
Timings: 7-9-7-24
vDIMM: 1.65 V
CPU Frequency: 4000 MHz
CPU Voltage (loaded): 1.41 V

Of course, everyone's mileage may vary.
 
Its not easy is it?

Do a 3000 CPU-NB now and watch how much easier it becomes stabilizing the memory.
 
It has been stable at 2800MHz CPU-NB with 1.2875v - should I push for 3GHz? I am using 1600MHz RAM...
 
no that was directed at Hokie.

With a 1600 i've found 2600 MHz to be enough.
 
Is there any benefit to me lowering to 2600 if 2800 is stable?

Other than saving on volts that is...
 
Hola,
Dolk thanks for creating “The Guild” (angels singing). It has saved me countless hours of testing and gave me a great starting point for my quest of OC’ing my phenom.

Long time viewer but first time poster. I’m in the process of OC’ing my new build and have run into a bit of a cross road. My parts list is:

Cpu: 955BE AM3
Mobo: Gigabyte GA-890FXA-UD5
Ram: G.SKILL Sniper 8GB (2 x 4GB) 240-Pin DDR3 SDRAM DDR3 1333
GPU: Asus 5850 Radeon.
PSU: Corsair 850W
On custom WC. In Win 7. Bios and Drivers are up to date.

1. Ran memtest 6 passes (8 hours) with no errors.
2. Under clocked ram to 1066 (mobo auto set timings ect.) didn’t want to stress ram.
3. Over clocked Cpu to 4.2 Ghz @ 1.475 volts (initially 1.450 with BSOD) using multi 200x21 and upped Cpu -NB as directed in Dolk’s guide to 2600 Mhz 200x13@ 1.25 volts. Stable at 24 hours of Prime 95 with small fft’s (temps maxed at 44 C). I did not change the HT Link (HTT) multiplier or HTT volts.
4. Upped ram back to stock 1333 (mobo reset timings ect.) and ran Prime 95 on blend. Prime 95 ran for 12 hours with no problems (temps never reached above 44 C). Sometime between 12 hours and 20 hours (while I was asleep) received BSOD.

I’m wondering what setting(s) should I change to get me stable at 4.2 clock. It seems like I’m almost stable since I ran Prime95 for 12 hours with no problems.
Should I up the HTT multi and HTT volts (if so what would you recommend)? Or should I up the DDR VTT Voltage Control since the only thing I changed was setting the ram back to stock? Or is it something else?

Any assistance would be greatly appreciated.
Thanks in advance
 
Back