AM2 X2 4000+ // CCBYF 0620UPMW...ocing report

[3oh6]

instead of clogging up that nice little 1M thread we have going, i will start this one here. i planned on just posting a thread after i have results but it seems like 30min prime stable 3GHz for an AM2 dual core is already a result so here goes...

my good buddy and personal CPU dealer, Dave over at Flickerdown (http://secure.flickerdown.com/) sent me a treat the other day and it showed up today!

processor in box // setup for testing // corsair 8500 c5s
http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/package-1sm.jpg (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/package-1.jpg) http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/setup-1sm.jpg (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/setup-1.jpg) http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/setup-2sm.jpg (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/setup-2.jpg)


August 23: Foxconn C51 gets a turn
Foxconn C51XEM2AA-8EKRS2H
OCZ2N9002GK - 2X1GB DDRII PC7200 4-4-3-15 2.1v (D9GKX) (http://www.xtremesystems.org/forums/showthread.php?t=106976)
CPU cooling Asetek Waterchill 3/8"
Silverstone Zeus 560W
images may be updated so be sure to refresh full size image.
280*10 = 2800MHz @ auto v (1.350v BIOS / 1.360v DMM) // DDR930 4-4-3-5 (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/2800_1_350v_prime-1.gif) - Dual StressPrime 12 Hours
290*10 = 2900MHz @ auto v (1.400v BIOS / 1.410v DMM) // DDR967 5-5-5-8 (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/2900_1_400v_prime-1.gif) - Dual StressPrime 6 Hours
300*10 = 3000MHz @ auto v (1.475v BIOS / 1.490v DMM) // DDR1000 5-5-5-8 (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/3000_1_475v_prime-1.gif) - Dual StressPrime 12 Hours

August 19: Last Updated
Asus M2N32-SLI Dlx Wifi
Corsair 2048-8500 C5s (review thread coming)
CPU cooling Asetek Waterchill 3/8"
TT Tough Power 700W
images may be updated so be sure to refresh full size image.
300*10 = 3000MHz @ 1.488v (1.49 BIOS / ??? DMM) // DDR1000 4-4-4-8 (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/3000_1_488v_32m-2.gif) - 32m SuperPi
300*10 = 3000MHz @ 1.472v (1.49 BIOS / ??? DMM) // DDR1000 4-4-4-8 (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/3000_1_472v_prime-1.gif) - 2hrs prime stable

310*10 = 3100MHz @ 1.66v (1.65 BIOS / ??? DMM) // DDR1032 4-4-4-8 (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/3100_1_66v_1m-1.gif) - 1m SuperPi
310*10 = 3100MHz @ 1.66v (1.65 BIOS / ??? DMM) // DDR1032 4-4-4-8 (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/3100_1_66v_32m-1.gif) - Dual 32m SuperPi
310*10 = 3100MHz @ 1.66v (1.65 BIOS / ??? DMM) // DDR1032 4-4-4-8 (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/3100_1_66v_prime-1.gif) - Dual StressPrime: 4 hours

3000MHz // DDR1200 4-5-4-5 2.5v (aka still more room) // SiSoft Sandra 12000 buffered (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/12000_sandra-1.gif)
3000MHz // DDR1000 4-4-4-8 2.1v // 4 hours crunching Rosetta stable (http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/3000_1_502v_ros-1.gif)

It appears this M2N32-SLI doesn't like to play with X2s that well. When priming, 32M, or any other CPU intesive stress test, the PWM shoots through the roof and around 2.9GHz it starts to reach 100C+ according to SpeedFan. At 3GHz DDR1200 I only get 10 mins of run time +/- 1 minute. I think PWM is the issue but may be the CPU as this memory with my single core is 48hr prime stable at 3GHz DDR1200 with only 2.4v and much tighter timings. The PWM only reaches 60Cish with the single core if I recall correctly.

The problem is, I can't find a hot spot anywhere on the motherboard. I haven't pulled the heatpipes yet but nowhere on the board is anything getting that hot that my fingers can find. Oh, and I'm calling it PWM but I have no idae what it is to be honest. I just assumed PWM with scaling temps like that. Either that or I have another memory controller that can't keep up with memory.

[andre X_X]

OMG, send me one....:D
how about pumping more voltage?? 1.55-1.6v maybe??

[3oh6]

:)...i have had it for 2 hours and it has been priming at 3GHz for about 1 hour and 45 minutes...i will fully explore the limits of this CPU over the next while. at 2 hours i will start going up again and see where i can get with 1.5v.

temperatures just don't seem to be an issue, even with the IHS. priming under 40C for the most part on the hotest core at 3GHz. looks like a good one for water but will prolly be crap when the temperature drops.

[andre X_X]

yeah, i saw many chips doing very great on air/water, but really sucks on subzero cooling....
but who cares, your doing very well... 3ghz 1.47v on Dual core - its like a dream came true...:D
maybe 3.1-3.15ghz would be the limit...

[Eldonko]

Awesome man, 4000+ FTW!

[afireinside]

Wow... I'm kicking my self over here. I was about to order one and decided to go for a 3800+ instead. I can't get past 2.85 :(

Grats on 12k sandra! What voltage is the ram at?

[3oh6]

thanx afi, i only paid marginally more than i would have for a 3800+ locally so this is by far the best CPU i have purchased in a while. it makes my 146 that runs 3GHz at 1.48 prime stable look like a joke. i mean, a dual core at pretty much the same voltage and frequency is solid. 13K will be done with this chip and the m2n32-sli...even if it takes a couple bags of ice ;). the question is, will the D9GMH do it without a VDIMM mod?

temps are awesome, this weak asetek kit has never looked like more of a champ but i will put the thermalright ultra-120 on it right away and it will be air stable at the same settings im betting...maybe a touch in VCORE will be needed.

what i am most happy about is the ability to run 2GB DDR1200 @ 3GHz. i can finally bench with a dual core! right now i am crunching Rosetta at 3GHz 1.472v, core temp is reading both cores around 37C ~ 39C. in a few hours i will see if it can run the D9GKX 4-4-3-5, DDR1100 @ 2.5v.

[afireinside]

I never tried 4-5-4-5 but if you can do that @ 600 you should do 4-4-3-5 at 550 no problem. Is the ddr1200 stable for anything other than sandra?

[3oh6]

yup, 32M, prime...the works i would imagine. i would be running rosetta at 1200 but i think i need the volts up at 2.4v to do that so for crunching it gets the 1000 divider and 4-4-4-8 @ 2.1v treatment. no point in wasting that voltage for crunching...save the Vs for benching where it belongs ;)

[DrJay]

I have the same cpu (CCBYF 0620 UPMW) on a Biostar Tforce 550. Been testing it for a couple weeks now. I was going to start a thread about it when I had enough data but have not gotten around to it yet....
Anyway, it ran stable at 3100Mhz on stock air cooling (stock for Opty dual s939) with 1.575 Vcore. Got it on water now...and with some decent memory so we'll see what happens. Tforce 550 needs a Vdimm mod though. Fortunately, one exists already :D . Good cpu's though. Congrats on picking one up. :toast:

[fakhrain]

If tf570u has a similar layout with tf550, vdimm mod should be quite easy...no solding is even needed. i just use pencil :D

[3oh6]

I have the same cpu (CCBYF 0620 UPMW) on a Biostar Tforce 550. Been testing it for a couple weeks now. I was going to start a thread about it when I had enough data but have not gotten around to it yet....
Anyway, it ran stable at 3100Mhz on stock air cooling (stock for Opty dual s939) with 1.575 Vcore. Got it on water now...and with some decent memory so we'll see what happens. Tforce 550 needs a Vdimm mod though. Fortunately, one exists already :D . Good cpu's though. Congrats on picking one up. :toast:
damn, yours is a good bit better than mine even. 3100MHz on air is one hefty accomplishment. obviously mine isn't stable at 3100MHz yet but it entirely could be the motherboard as that PWM can't be stable running that hot. i'm not sure i trust the reading but the temps scale with memory/CPU scaling. i just wish i could find the hotspot on this thing to cool it down. feel free to high-jack this thread at anytime with your results...maybe try to steal the first post of the scond page (should have reserved the second reply).

[DrJay]

damn, yours is a good bit better than mine even. 3100MHz on air is one hefty accomplishment. obviously mine isn't stable at 3100MHz yet but it entirely could be the motherboard as that PWM can't be stable running that hot. i'm not sure i trust the reading but the temps scale with memory/CPU scaling. i just wish i could find the hotspot on this thing to cool it down. feel free to high-jack this thread at anytime with your results...maybe try to steal the first post of the scond page (should have reserved the second reply).

Water hasn't helped the overclock. Probably running a little cooler though. (Can't br sure, the Biostar T-utility is way off)
Thanks for allowing further hijacking. :toast:
Is this CCBYF seen on any other AM2 CPUs?

EDIT: The Vcore in the attachment is set to 1.575,...T-utility reads 1.58.

[3oh6]

Water hasn't helped the overclock. Probably running a little cooler though. (Can't br sure, the Biostar T-utility is way off)
Thanks for allowing further hijacking. :toast:
Is this CCBYF seen on any other AM2 CPUs?

EDIT: The Vcore in the attachment is set to 1.575,...T-utility reads 1.58.
haha, let the high-jacking commence! it's all relevent IMO...

i see your single 8M (5m 7s) and raise you dual 8M (1/core) at 3100MHz in 5m 4s.

http://img.photobucket.com/albums/v220/3oh6/xsforum/ccbye_x2_3800/metoo-2.gif

I also updated the first post with the reason for no updates. It took me all of yesterday to figure the issue out as to why I can't get stable above 3GHz or at 3GHz with certain memory speeds.

[3oh6]

Is the ddr1200 stable for anything other than sandra?
i lied to you...i thought i ran 32M at DDR1200 but don't have screens so it didn't happen. check the bottom of my first post for an explanation...i think it is this motherboard. i am going to dust off the C51, i also have this biostar 570 ultra board here that might be just thing i need (very basic, single chipset, easy to cool).

i think that is also the reason my 3100 prime shuts the system down at the 10 min mark. coincidence that 3GHz 32M DDR1200 shuts the system off at 10 mins and 3.1GHz StressPrime DDR775 shuts off at 10 mins too with that one temp over 127C??? didn't think so...this chip has plenty left in it ;)

[3oh6]

just posted:
3100GHz @ 1.66v // DDR1032 @ 4-4-4-8 w/2.5v - Dual 32M (this is the AMD forum after all)
3100GHz @ 1.66v // DDR1032 @ 4-4-4-8 w/2.5v - 1M
in the first thread. PWM only gets to 115C under load so it looks like this CPU can at least do 3.1GHz stable before I even start to chill the water. Load temps are a paltry 42C~44C and I cannot stress enough how weak this water cooling is unless I am on a bare core and these temps are pretty much identical to what I get with a naked 3200+ at the same frequency and voltage...this chip is ice cold and the IHS has to be decent.

i will get a heatpipe cooler on it and see if 3GHz is stable at 1.5v or so...after i'm done playing with memory of course.

[DrJay]

Nice job 3oh6!:clap:

[dave_graham]

remember, this is the SAME thing as the Opteron 1210 (the 4000+ was EOL'd and replaced by the 1210), so...depending on how batches track, you could expect similar results from those shipping.

dave

[knob]

nice job mate. Well done!

[3oh6]

the 4000+ has been moved over to the Foxconn C51 and taken its cooling with it. sorry but the only air cooler i have that mounts AM2 at the moment is busy so unless i can get an AM2 Ultra-120 hold down...it will be water results for a little while yet.

2800MHz @ auto v (1.35v BIOS / 1.36v DMM) // DDR930 4-4-3-5 - just added the first Foxconn result

[hypocriTe]

Just a quick question but,
Was wondering how much effect the bigger L2 cache has in overall performance? :confused:

Very nice clocks btw. :toast:

[CWell1337]

Hey guys, I know this thread has been dead for awhile but I just googled my stepping and this thread was the first hit. Hoping the OP can give me some advice!

Anyways, here's my rig:
AM2 AMD Athlon 64 X2 4000+ (BBBYF 0620UPMW)
ASUS M2N32-SLI Deluxe
2GB Corsair XMS2 PC8500C5
Zalman 9700
BFG 650W PSU
eVGA 7900GTX
160GB WD SATA2
320GB Seagate SATA2
Sony Lightscribe DVD+/-RW
Lian-Li PC-65 Plus II

So at one point I had the same CPU, motherboard, and RAM as you were testing! As you can see I'm on air, but I'm getting a pretty good OC on here. I've run dual Prime95 for 24 hours on 2.9Ghz on 1.552v vcore (dropped to 1.536v on load). Right now I'm dual Prime95ing on 3.0Ghz at 1.57v vcore.

Here's a SS of my current CPU-Z. Any suggestions on RAM timings/speed/voltage for me? Right now I'm at DDR2-600 3-3-3-7-1T on 2.3V. Can I take the voltage much higher? Any suggestions in general? I would love to be able to get down to the 26s SuperPi 1M time you were getting.

Hoping to hear back from you! Thanks.

C

Also, any idea why CPU-Z is showing a vcore of 1.3v which I know I set in BIOS to ~1.57? You can see the Everest reading in the SS I added.

[3oh6]

well, you basically have the same setup so you should be able to reach similar results with the memory. DDR600 3-3-3-7, even 1T isn't as fast as DDR1000 4-4-4-8 2T i don't think.

start with seeing if you can run stable with the memory on the /6 divider for DDR1000 and set the timings to 4-4-4-8 2T with a reasonable amount of voltage. you can use the secondary timings from that 1M SS in the first post. personally, i wouldn't run DDRII over 2.3v in a 24/7 machine...even with a fan actively cooling them. so staying at 2.3v or less is the best idea.

also keep in mind that the 1M is also at 3100MHz and SPi isn't everything. at 3GHz with memory at DDR1000 4-4-4-8-11 2T, SPi 1M should be in the 27s range on a tweaked OS, 27.5s on a standard install. HTH

[CWell1337]

Alright, so after this Prime95 is done running at 3.0Ghz I will memtest86 my memory at DDR2-100 on a /6 divider (that the DDR2-667 divider or the DDR2-533 divider?) at 4-4-4-8-2T on 2.3V. You think that should work on a 300Mhz FSB?

I don't know if I plan on running at 3.0Ghz 24/7 even if it's stable. 1.57v vcore is too much I think. 2.8Ghz I can get stable at 1.4v vcore so that will be my 24/7 running speed. Just trying to OC to 3.0Ghz and better RAM performance for max benchmarking in SuperPI, 3DMark etc.

Thanks for the fast reply! I'll update tomorrow night, once this Prime95 has a chance to run for 24 hours and I try the RAM tweaks.

[3oh6]

yeah, those settings sound good, although, you might not need that much voltage for the memory but start there and then work your way down. for DDR1000 @ 3GHz yes it is the 667 divider. 667 = CPU Freq/6 and 533 = CPU Freq/8. there is a great chart someone made posted here at XS...

http://www.xtremesystems.org/forums/showthread.php?t=105798&highlight=am2+divider

it has all the AM2 dividers listed for each CPU multiplier and DDR speeds at given CPU frequencies. very helpful until you figure out what they all are.

[Noob-ftw]

Wow, great results from the hidden AM2 series.

[CWell1337]

Hmm, I got anxious and decided to try those memory timings out now on my 2.9Ghz OC. Well it seems that I can't successfully use ANY OTHER configurations. I tried 2.4v at DDR2-667 on 4-4-4-12-2T and I couldn't stay in windows... what the hell. I knew there was a reason I had my timings like they are...

Yeah not much else is working. DDR2-667 5-5-5-15-2T wont work on 2.3v. Might I have bad RAM? It gets great timings on a 1:1 ratio, but I can't get :banana::banana::banana::banana: to run on a divider.

[CWell1337]

What do you think is my problem? Why can't I run my RAM at higher speeds? Do I have a BIOS setting wrong? or maybe a hardware issue?

[3oh6]

nothing is gauranteed in overclocking and certainly not with RAM because there are so many factors in play. also, there is no 1:1 with AM2, everything is a divider so it isn't a divider issue i wouldn't imagine.

i don't have this motherboard anymore so i am not sure i could provide much help even with proper screenshots from you but if you want to post some CPU-Z and SysTool screenshots of the memory timings that are not stable, i could possibly offer some advice on changes.

[Solarfall]

nicelly done mate, now try 3.2ghz or more

[CWell1337]

The two screen shots are stable boots into windows. The following are configurations I tried unsuccessfully. All were done on 2.3v.

DDR2-667
4-4-4-12-2T
POST Error

DDR2-667
5-5-5-15-2T
POST Error

DDR2-533
4-4-4-8-1T
Windows Crash

DDR2-533
4-4-4-10-1T
Windows Startup Crash

I don't understand why with the exception to the two SS I attached, I can't run my RAM anywhere else. Do you want BIOS screen shots of what everything looks like? Maybe it's something else I need to change to get the RAM to run stable at higher speeds?

Thanks in advance,

C

[3oh6]

it sounds like you should be contacting Corsair tech support because there may be something worng with your modules. if they are 8500 C5s, they should run be running DDR1066 at 5-5-5-X-2T with 2.2v or even less. sorry i couldn't be more help.

[CWell1337]

Thanks anyways, off to Corsair HOH I go.