The Phenom II List of Overclocks

[FlanK3r]

good job

[pcnazz]

Hi guys . I use Prime95 a lot . I have seen it crash just under 1 hr . I read this whole thread and a lot of people have issues on running Prime95 for more than 1 hr . Why don't we use OCCT . It is a 1 hr run and it is a very good stability test . http://http://downloads.guru3d.com/OCCT-%28OverClock-Checking-Tool%29-3.0.0-download-1880.html

[Lunait]

I see that many of you guys leave your HT clock at 2 GHz, even though your NB is overclocked. Is there any special reason for this?

[pcnazz]

HI Lunait . I run my PH II X4 940 @ 3.913 MHz , HT @ 2608.9 MHz , NB @ 2809.5 MHz . Its good to keep HT and NB as close as possible to eliminate any bottle necking . Its important to know I am overclocking all three by multiplier's .

[Lunait]

HI Lunait . I run my PH II X4 940 @ 3.913 MHz , HT @ 2208.9 MHz , NB @ 2809.5 MHz . Its good to keep HT and NB as close as possible to eliminate any bottle necking . Its important to know I am overclocking all three by multiplier's .

Yeah, but isn't HT supposed to be lower than or equal to NB? If so, why not run HT at the same speed as the NB? Or does that increase the risk of instability?

[mursaat]

@Lunait: reading a bit I found HT have almost no effect in performance, while NB oc lowers L3 and memory latency, which is always welcomed. That I guess is what makes people leave HT at 2ghz and only overclock the NB.

//EDIT: I've always overclocked through FSB speed rather than multiplier (mostly because up until recently all my procs were multiplier locked). I mean, it's always been better to OC the FSB leaving a lower multiplier if possible, why suddenly it's not that important? I've seen lots and lots of OC with 200x19.5 instead of 279x14. New board doesn't support high FSB or what?

[Lunait]

@Lunait: reading a bit I found HT have almost no effect in performance, while NB oc lowers L3 and memory latency, which is always welcomed. That I guess is what makes people leave HT at 2ghz and only overclock the NB.

//EDIT: I've always overclocked through FSB speed rather than multiplier (mostly because up until recently all my procs were multiplier locked). I mean, it's always been better to OC the FSB leaving a lower multiplier if possible, why suddenly it's not that important? I've seen lots and lots of OC with 200x19.5 instead of 279x14. New board doesn't support high FSB or what?

Well, that's the good thing about AMD Black Edition CPU:s - you can overclock each component separately using multipliers (although the RAM has to be overclocked using FSB). Or are you saying that overclocking using FSB is better than using the different multipliers, even though the same frequencies are achieved?

[mav2000]

Agree with you there...I would rather OC the CPU, NB and to some extent the HT separately using the multis and then use the fsb for the smaller increments that can take u to the best highest clock...

[Ben_Brown]

Overclocking by way of the 'FSB' does not always yield the best results. If you take a look at the two screens below for the example you will see that the first one is using multi's only to overclock with the CPU at 4013MHz and the NB at 2809MHz and the score was 16696. After that I tried to overclock using the reference clock (FSB) and the second screen shows the CPU at 4005MHz and the NB at 2937MHz but the score is about 500 points lower. In both these test memory is almost the exact same speed. You would think with the NB aver 100MHz higher that the score would be higher as well.

[mursaat]

That was my question in the first time, in the 939 ol'times () it was all about getting a DFI SLI-DR to get ridiculous FSB speeds to use with a 9x multiplier locked Opteron 165. IIRC there were FSB speed records like 600+mhz (being 200mhz the default speed too).

[pcnazz]

Yeah, but isn't HT supposed to be lower than or equal to NB? If so, why not run HT at the same speed as the NB? Or does that increase the risk of instability?

Hi Lunait . I'm apologize for the error I typed in my message . My HT link is at 2608.9 MHz , NB 2809.4 MHz . Yes that was my initial point I like to run both HT and NB at the same speed if possible . My memory runs at 535.2 MHz ( 1070.4 MHz ) I fixed the error in the original post .

[mursaat]

Ok after 2 days of testing, I got some numbers. No screenshot as I am in my laptop, you'll have to trust me passes 1h OCCT:

Phenom II 550 unlocked X4 with Asrock 790GMH/128M with OCZ Reaper DDR2-800 (really weak ram, so not much ram testing), PC P&C Silencer 370w.

1.175vcore, 12.5x260, 3.25ghz
1.3vcore, 13x260, 3.51ghz
1.4vcore, 18,5x200, 3,7ghz (havent tried high FSB with this vcore yet)

Not going much further, as I'm on the stock AMD X2 cooler, it's already 68ºC when stressing at 1.4v

I'm gonna settle for 3.25ghz with 1.175vcore, having the cheapest PII running closer to the most expensiveone is sweet enough for my mom's computer

Either that or 3.5ghz with stock vcore, what would you say?

[Zeus]

955BE 0915APAW watercooled.

Time sucks but it did 32M at 4.1GHz non the less.

[dragonwolf13]

Hi guys new here I thought I post my OC to start with.
I did this with air 3515Mhz on a Gigabyte MA790FXT-UD5P with a Phenom II X3 720BE, Visat64, wish I could get it higher but can't get it stable enough to do it.

[blackwarriors]

Hi guys, my 955 BE in aircooling with Thermalright True Copper + 2 Nanoxia 2000rpm





Minimum vcore for boot



http://valid.canardpc.com/show_oc.php?id=690661

[Zeus]

That is a damn fine cpu you've got there blackwarriors!

[ridney]

wow! that's an amazing chip blackwarrior! 1st time i have seen a 955 at 4ghz with such a low voltage, congrats!

[FlanK3r]

Hi guys, my 955 BE in aircooling with Thermalright True Copper + 2 Nanoxia 2000rpm





Minimum vcore for boot



http://valid.canardpc.com/show_oc.php?id=690661

cool...try, for us, max validation with 1.475-1.55V. Think 4400 MHz is possible (maybe 4500-4600 MHz)

[Muraretu]

I have 2 OC Cpu.
This image:


and this validaition:
http://valid.canardpc.com/show_oc.php?id=634197

[Darknight]

Max Validation ( Not Validate)

http://valid.canardpc.com/show_oc.php?id=700571

Validation (Had to 5Ghz, but became the 5.1)

http://valid.canardpc.com/show_oc.php?id=700252

Super Pi 1Mb



Not tested for RAM to 6-6-6-15, I did not have time

Max HTT and NB



Not tried it, had no more DICE, nor was missing in best conditions

*Sorry My english

[Evil]

Hi guys, my 955 BE in aircooling with Thermalright True Copper + 2 Nanoxia 2000rpm

Damn, golden chip. What´s the fpo?
Getting 955BE CACYC AC 0918BPAW next week, anyone have any info on that?

[Zeus]

Can anyone explain why this cpu has a stock Vcore of 1.35V?



This is even with a 64 bit OS which is supposed to be harder on the cpu stability, ie should require more vcore for stability.

Don´t know if i can replicate this with the stock cooler though, temps were only 34ºC full load, idle only dropped 1-2C° coming from 1.38Vcore.

[equinox1]

pretty happy with these, 4305mhz on water for spi 1m. pretty decent 32m also

[Trent04]

Validation
On crappy water

[BiGSpender]

Can't get past 3.7 stable on my x4 955 you can see my case and cooling in my sig. 3.8 crashes hard as soon as I start pime95 or occt.

Thanks for your help