The Phenom II List of Overclocks

[demonkevy666]

my chip idles at 40C at 1.488 now,load has only been up 54C. I can use 18.5 and above multiplier. I can't get above that really yet. I tried 19 got into windows, tried 19.5 could get into windows went back down to 19 and could get in to windows again. is this because I left my chip OC at 3.2ghz and then came back like 6 hours later, played games and other things. seem the longer I sit on one multiplier the higher I'll get next time

[keithlm]

For a clock of 3.6Ghz: Currently at 1.47V on the core. In the bios I set this voltage and a multiplier of x18. I have also set the NB to 2.4Ghz in the bios. (The HT stays at 1.8Ghz... I can't get the bios to change it correctly.)

For a clock of 3.713Ghz: Using AOD I can also increase the Bus speed from 200 to 206. (No voltage change needed.)

I have run Prime95 for many hours at 3.6 (Just the x18 multiplier) but have not yet done it with the added 6Mhz to the bus speed. (I'll do that later tonight.) UPDATE: After running 6 hours Prime95 crashed at 3.7Ghz.

But I also got some nice Futuremarks scores: (Using 2x4850 video.)


3DMark Vantage: http://service.futuremark.com/compare?3dmv=690568
P13504
CPU 12724
GPU 13785

3DMark06: http://service.futuremark.com/compare?3dm06=9635306
20031 <<<--- OH YEAH that made me happy. (I would have cried if I got like 19,950)
SM 2.0 7895
SM 3.0 9892
CPU 5131

PCMark05: http://service.futuremark.com/compare?pcm05=1821014
11142

EDIT: Interesting that the next post by pershoot has almost the same settings. Seeing his results: I might be able to decrease my voltage after I test for stability. Although he got 2.6Ghz on the NB. I can't.

[pershoot]

final clock until i can sort some heat issues at higher volts.

http://img243.imageshack.us/img243/2...henomiiyu4.jpg

[Sgt.McRuff]

Pershoot, I got a question: Can you watch vcore and see if it drops a bit under high load?
I believe the 790FXB-M2RSH(bios?) has an error under high load with Ph II. Just want to hear it from multiple people to get rid of any doubt that its something else.

[crazydiamond]

mcruff, seems most people have been experiencing that regardless of board

[Sgt.McRuff]

mcruff, seems most people have been experiencing that regardless of board

Im not to sure about that. Went threw some amount of pages and didn't see anyone else complaining about it with other boards. And other peoples occt vcore graph looked normal. Its not a spike but a constant thing.

Some examples:
http://i380.photobucket.com/albums/o...2h04-VCore.png
http://i380.photobucket.com/albums/o...4h40-VCore.png

Can you link me to posts with people having same problem with different boards? Would appreciate it.

[crazydiamond]

i just member reading people mention it a few times cuz i thought it was strange since phenom 1 is opposite... heres a pic from earlier in thread describing what i think your talkin about..

edit: lol looks like the same board, i didnt rip the pic from one of your posts did i?

[Dumo]

With ss...

[Morais]

Nice one dumo.What are the real temps and how is the condensation with a single stage?

...

Dang, time for get rid of these old x2 stickies and put this thread on it!

[Titan7171]

final clock until i can sort some heat issues at higher volts.

http://img243.imageshack.us/img243/2...henomiiyu4.jpg

Dont know if you noticed but your 12v and 5v line look low.

[pershoot]

Pershoot, I got a question: Can you watch vcore and see if it drops a bit under high load?
I believe the 790FXB-M2RSH(bios?) has an error under high load with Ph II. Just want to hear it from multiple people to get rid of any doubt that its something else.

ya it drops a bit. for ex: 1.42v idle and 1.41 load. had the same thing happen on my phenom I as well. same thing happened on the 790fx-m2r.

[pershoot]

Dont know if you noticed but your 12v and 5v line look low.

it should be fine. i dont put much faith in the motherboard sensors/software readout of the voltage rails. the PSU is more then adequate.

[Particle]

Here is a bit of scaling information for Phenom II in that definition-of-synthetic benchmark called Sandra 2009 SP2. I have to wonder if these results would be substantially better on another board. As I recall, when I went from my K9A2 Plat or M3A78-T to my A79A-S, my score in IBT went down by like 3-5 GFLOPS at the same clock speed using my 9850. It's hard to remember which board it was since I've blown through so many due to FETs giving up the ghost at lame voltages like 1.40V.

Anyway, you'll notice that scaling is between 99% and 101% for each jump, so in this synthetic test we're seeing linear scaling pretty much.

[Ben_Brown]

Well I tried to copy Oldguy932 and disable 2 of the cores to try and get a higher clock. I don't think it really helped much on my water setup. Below are the results. My next step is to try HT at 300 to see if that makes any difference.

[chew*]

Looks like it time for me to join the fray, week 48 chip Hopefully my local computer shop has the gigabyte GX board in stock, Don't really have money for another vid card atm.

[Rammsteiner]

WTF, NB is a to OC though

Changed multi to 9, it BSOD'd right after POST.
Changed Vnb to 1B from 1C, BSOD'd right after boot.
Changed Vnb to 1A from 1B, Error'd in Memtest quite fast.
Changed Vnb to 19 from 1A, Error'd in Memtest as I browsed the net.
Changed Vnb to 18 from 19, running now for 30m+ with browsing the net

But it takes quite a bit to find out what it wants Also I noticed NB-clocks/voltage have quite an impact on CPU temp Not like 10C, but an easy 2~3C rise under load.

Im about to think to up the HTT one notch to 293. CPU is now at 3796Mhz, not that magic 3.8Ghz But first testing this, will run Memtest for another ~20 minutes, then a 10 run Intelburntest x64. From there Ill try to raise the HTT from 292 to 293 and test again. After that... my biggest bottleneck... Vista Although I like Aero a lot, Im already thinking to change the looks to windows classic

And of course trying to fix the firmware of my RAID controller

[Rammsteiner]

Why is it always me having retarded IMC's

I tried Vnb all the way upto '14' (1.3V) but it refuses not to BSOD under Intelburntest I tried different HT multipliers, I tried to use Drive Strength Weak and Normal mode, it just always BSOD's.

Now Im going to run on previous stable OC and test that just to be sure it's indeed stable, although I hope otherwise it would probably mean something is dieing

PS, is there any thing I might be forgetting? AFAIK NB OCíng is just like CPU? Add Voltage if higher clock isnt stable? Although it doesnt work, so I prolly hit a wall then?

[Throwed]

Looks like it time for me to join the fray, week 48 chip Hopefully my local computer shop has the gigabyte GX board in stock, Don't really have money for another vid card atm.

I hate you! I hope you can't find the board you want either......SOB

[Ben_Brown]

Why is it always me having retarded IMC's

I tried Vnb all the way upto '14' (1.3V) but it refuses not to BSOD under Intelburntest I tried different HT multipliers, I tried to use Drive Strength Weak and Normal mode, it just always BSOD's.

Now Im going to run on previous stable OC and test that just to be sure it's indeed stable, although I hope otherwise it would probably mean something is dieing

PS, is there any thing I might be forgetting? AFAIK NB OCíng is just like CPU? Add Voltage if higher clock isnt stable? Although it doesnt work, so I prolly hit a wall then?

How high are you trying to overclock the NB? The reason I ask is I need 1.4250 for 2800MHz and 1.4750 for 3000MHz for voltage.

[jonspd]

NB voltage should only help with overclocking the htref or ht link speed the nb speed is on the cpu die and should be able to be reach around 3ghz IMO could be wrong about this.

[Rammsteiner]

How high are you trying to overclock the NB? The reason I ask is I need 1.4250 for 2800MHz and 1.4750 for 3000MHz for voltage.

Something reasonable to match my DDR1160+ RAM. The higher the NB the more efficient the RAM gets as well. But I see people hitting damn high clocks on the NB, although most are WC'd or even sub-zero, but still. Read on further this post cause I made a up

NB voltage should only help with overclocking the htref or ht link speed the nb speed is on the cpu die and should be able to be reach around 3ghz IMO could be wrong about this.

Yeah, we already saw 3Ghz happening, but that was with 1.45~1.5V. And to be honest, Ive no single idea how safe those Voltages are. In the end an IMC ain't a CPU core, but whether it's more sensitive or maybe not at all, I dont know.

Anyway, I just ran my previous 'stable' OC test and it BSOD'd. Since Im writing everything down on paper I decided to look at it again, turning out I only tested my NB with a 8x multi (2.33Ghz) using Memtest for ~8.5% coverage Now that's a stupid screw up since that doesnt say anything at all. Already found it strange I could go from 7x to 8x using the same Vnb and after that no scaling anymore.

Anyway, so Ill have to test 8x again Does anyone have a clue what might be moderate safe (FWIW) Voltages for the NB though?

[demonkevy666]

Something reasonable to match my DDR1160+ RAM. The higher the NB the more efficient the RAM gets as well. But I see people hitting damn high clocks on the NB, although most are WC'd or even sub-zero, but still. Read on further this post cause I made a up


Yeah, we already saw 3Ghz happening, but that was with 1.45~1.5V. And to be honest, Ive no single idea how safe those Voltages are. In the end an IMC ain't a CPU core, but whether it's more sensitive or maybe not at all, I dont know.

Anyway, I just ran my previous 'stable' OC test and it BSOD'd. Since Im writing everything down on paper I decided to look at it again, turning out I only tested my NB with a 8x multi (2.33Ghz) using Memtest for ~8.5% coverage Now that's a stupid screw up since that doesnt say anything at all. Already found it strange I could go from 7x to 8x using the same Vnb and after that no scaling anymore.

Anyway, so Ill have to test 8x again Does anyone have a clue what might be moderate safe (FWIW) Voltages for the NB though?

yeah I think since we going for higher core speeds the NB is just going to add to temperatures on air cooling I just tied 2.6ghz with 3.755ghz. it wasn't stable but it got into windows.core votls and nb volts where the same 1.4875. 43C from core temp, from bios read this OC at 37C core temp is showing 41C idle

chips really like water cooling.

[Oese]

@ oldguy: could you correct my entry? its water not air..

and another one here:

Oese
Phenom II 920
CACVC AC 0849BPEW
Batch*******0540
DFI LP UT 790FX-M2R
4x1GB Ballistix D9GMH
Tagan 2-Force II 700w
AMD 4850 800/1025
Watercooled
XP 32bit

[Ben_Brown]

The NB is just as important if not more important that the CPU speed. A good example would be the following two links

These are Super Pi 1m runs.

This one is with CPU @ 5000MHz and a NB @ 1000Mhz
http://www.hwbot.org/compare.do?resultId=809287

This next one is with CPU @ 4231MHz and NB @ 3000MHz
http://www.hwbot.org/compare.do?resultId=811581

Which of the above do you think has the higher run....

[Lastviking]

The NB is just as important if not more important that the CPU speed. A good example would be the following two links

These are Super Pi 1m runs.

This one is with CPU @ 5000MHz and a NB @ 1000Mhz
http://www.hwbot.org/compare.do?resultId=809287

This next one is with CPU @ 4231MHz and NB @ 3000MHz
http://www.hwbot.org/compare.do?resultId=811581

Which of the above do you think has the higher run....

Hmm same windows? , i dont think its fair compair.....its easyer to compair if you do the compair by your self.