Attn 939 X2-3800+/4200+ owners!

[avi85]

How high does your 939 X2-3800+/4200+ go on stock vCore

Please don't answer the poll if you have a different processor, thanks.

 

[myocardia]

You need to specify which chips you're talking about, I would think. I hear that AM2's are going quite a bit higher with pretty close to stock voltage, which would mean that they're going higher on bone-stock voltages, too.

edit: I can almost guarantee you that if you're talking about Skt. 939's, it's going to be very close to 2.40, assuming they tell you the truth.

 

[avi85]

S939, thanks for the heads up

 

[formulav8]

I can get right at 2.6ghz on stock voltage. I did have to slightly raise it a couple months ago. It seems the older it got, the more it needed voltage at that 2.6ghz speed. Could just be a fluke though. I have tried many times though to get it to 2.7ghz, and no matter what it wouldn't be stable even feeding it some extra voltage. My chip is over a year old too. It was made around August 05. AMD should have better yields then when I got mine.

Also, AMD WILL be using DSL silicone on the s939 chips. That should give them alittle higher oc potential. But the problem is, AMD said they will not make any anouncement or anything when that transition takes place. It somewhat appears if the cpu part number has CU, it may be DSL Silicon. I've seen a few single core s939 chips with those part numbers. All of the AM2 cpu's have that part number section higher than CD (CD was the Toledo (Venice Based) core). They are mainly CU based part numbers. And ALL AM2 cpu's are using DSL silicon.

My 2 pennys anyways



Jason

 

[Pederv]

About 2270MHz (10X227).

 

[stevty2889]

I'ts been a while, but it was somewhere around 2.3-2.4ghz before I had to raise the vcore. Been running at 2.618ghz 1.475v for a while now.

 

[Zebo]

2800Mhz stock vcore stock HSF. But thats not normal from what other people say..But both my 3800 939 chips do it.

My 4200+ was a second week one from last year, so not so good only 2450Mhz.

 

[Mucker]

X2 3800+ @ 2620 MHz @ 1.58 volts.....Epox 9NPA+ Sli....TT Big Typhoon @ 57C max load both cores in 80F ambient air....

2620 MHz

mWave purchased
LDBFE 0622 WPMW
E4 Manchester

Does better on my DFI Ultra-D. It's not memory controller strong like Opteron, but nonetheless a pretty good dual core for $150, equivalent to a stock x2 4800+

 

[Dave Sz]

Amd stopped making or will stop making 939 chips so why would they add a new material to them?

 

[dawgfan6]

Who-hoo!! I just got my x2 3800 939 w/ ASROCK 939dual bios 2.2 (reusing my 6800GT AGP and ram...most cost effective upgrade for me - $157 delivered). I am at 2750 "stable" out of the box w/ stock heatsink and fan (aluminum...not the heatpipe) and stock voltage (I'm not sure you can up the voltage on this MB w/ X2 without a Vmod, I'm at 1.4). I haven't really gotten around to tweakin' the details yet, but I crashed 3dmark06 at 2.8 (booted into and browsed around in windows at 2.8 without a problem...machine didn't crash, just 3dmark). I just started at 10x240 and bumped 10 at a time until the program crashed at 2.8, then I dialed down to 2.75 and it's been smooth sailing. I couldn't be happier!

 

[avi85]

c'mon people, lots of you have at least one of these procs!

 

[itsrunninghot]

I have my LDBFE 0619WPMW at 2.6 using 1.52 vcore. I'm thinking about running 1.6 vcore to try and get 2.7-2.8 stable. I almost dont care if it dies because it will give me a reason to do a new build.. maybe conroe once they get more reviews on retail parts. So how long should this chip last at 1.6vcore assuming load temps dont exceed 50-55c?

 

[Sparky19692]

Zebo where did you get GO-Orthos?
2.5 10X250 max at stock Voltage. Temp Case 30c CPU 51c

 

[shamans]

Originally posted by: avi85
c'mon people, lots of you have at least one of these procs!

Mine's gonna arrive in about a week
Will let you know then.

 

[videopho]

My is x2 3800+ hits a brick wall @ 2.32ghz at stocked Vcore.

 

[formulav8]

video, did you lower your htt multiplier to a 3x/600 mhz? Someone mentioned to you in the other thread to do that. Also, set the memory divider to 100mhz. Then slowly raise the htt mhz and see what you can get.

IF you did those things already, then just ignore my post



Jason

 

[BlueWeasel]

My 0612UPCW 4200+ runs 2.5Ghz (250x10) at 1.40-1.42v, and I'd guess it could probably hit 2.4Ghz at 1.35v. Running a modded ASrock 939Dual-SATA2, the next step up in voltage is 1.52v and it appears to be stable at 2.6Ghz.

I've also tried a 0613VPAW and it overclocks slightly less than the UPCW.

I was kinda hoping for more out of the 4200+ and was thinking 2.7Ghz at 1.50v. I should've just went with the 3800+

 

[chazzzer]

My new 3800+ hits at least 2.45 at stock voltage (that's 1.375v according to CPU-Z). I test with two copies of Prime95, each locked to a specific core. At 2.5GHz core 0 runs fine but core 1 will error-out within about 10 minutes. At 2.48 core 1 lasts about an hour. It's 24 hours stable at 2.45. This is all with the stock cooler for now, running on an ASRock 939Dual-SATA2 motherboard.

So, 2.46 or 2.47 is probably it. I'm done poking at it, since I'll be doing the VID3 volt mod on the mobo soon. No sense wasting time finding what's the absolute max performance at stock when I won't be at stock for long.

Even if I can never get over 2.45, it was a great deal for $150. I've been wanting an A64 X2 since they were first announced, but couldn't justify the price until the recent drops. I've had a long line of dual-proc computers, starting with a dual P3-450MHz (currently running Linux, when I turn it on) and going through several different boxes along the way (still have a dual P3-966MHz general purpose machine and a dual P3-1.13GHz RAID server).

I'm happy that dual-core has finally brought the responsiveness of dual-proc computing to the masses, and at reasonable prices now.

 

[Dravic]

3800+ X2 LCBIE 0619RPMW

First shot was 1.45 at 2.75 loads windows fine and played wow for about 2 hrs no problems, but not dual prime stable.

Prime stable at 2.7 at 1.5 volts, and currently at 2.65 @ 1.475 volts


I can even boot into windows at 2.8 @ 1.5 but CPU 1 bails out of prime immediately

Core 0 is stable at 2.75 @ 1.5 volts, but core 1 errors out of prime.

Anyone think I?ll get better results with a better HSF; I?m currently using a Zalman CNPS7000B-AlCu LED , I?m not to sure since temps at prime stable are 36-37c idle, and 46-48c load are pretty good for 1.475 and 1.5 volts.

 

[PianoMan]

Ummm... haven't tried to overclock... yet - and I've got a DFI board.

PM

 

[Dravic]

Originally posted by: Dravic
3800+ X2 LCBIE 0619RPMW

First shot was 1.45 at 2.75 loads windows fine and played wow for about 2 hrs no problems, but not dual prime stable.

Prime stable at 2.7 at 1.5 volts, and currently at 2.65 @ 1.475 volts


I can even boot into windows at 2.8 @ 1.5 but CPU 1 bails out of prime immediately

Core 0 is stable at 2.75 @ 1.5 volts, but core 1 errors out of prime.

Anyone think I?ll get better results with a better HSF; I?m currently using a Zalman CNPS7000B-AlCu LED , I?m not to sure since temps at prime stable are 36-37c idle, and 46-48c load are pretty good for 1.475 and 1.5 volts.

yes i quoted myself...


quick update

The bios and mobo monitoring temps were WAY off, i started using Nvidia's monitoring tool (which is good for monitoring and should be unistalled directlty after as the ntune service can effect network latency [at least on xp 64bit].


I decided to back it down to 2.6 GHz at 1.45 volts because I found out @ 2.7 GHz at 1.5v I was idling at 38c similar to before, but at load it was running at 65c OUCH, a 19c difference in reported temps.

Ordered a scythe mine cooler with a 120mm fan instead of the 100mm silent fan that comes with it, as silence is not a worry for me.

Temps now at 2.77ghz prime stable @ 1.475v as per nvidia?s monitoring tool, 36c at idle, and 52c (yesterday with no AC on) load. I?ll have to check again now that the AC is working again in the house, but I?m assuming it will drop 2-3c. almost stable at 2.8 @1.5 v but the second core errors out after a while in prime.

Now I?ll probably back it down to 2.75ish and live with it there, as I like to find the limit and back it off just a bit to ensure for temp spikes and the like.

The scythe mine cooler if prett sweet. And much lighter then the ninja or TT 120. 560 grams as compared to 665 for the ninja and 900+ for the Tuniq Tower 120.

good air cooling round up - summary page

start of the roundup

 

[PingSpike]

I dual primed overnight at 2.5ghz on my 3800+ x2 with stock vcore and then just set it on 2.4ghz for daily use.

It was running pretty hot by the end of it, but I have a arctic cooling freezer pro 64 and I think the temp sensors on my board may be weird.

 

[imported_inspire]

Originally posted by: Dravic

Originally posted by: Dravic
3800+ X2 LCBIE 0619RPMW

First shot was 1.45 at 2.75 loads windows fine and played wow for about 2 hrs no problems, but not dual prime stable.

Prime stable at 2.7 at 1.5 volts, and currently at 2.65 @ 1.475 volts


I can even boot into windows at 2.8 @ 1.5 but CPU 1 bails out of prime immediately

Core 0 is stable at 2.75 @ 1.5 volts, but core 1 errors out of prime.

Anyone think I?ll get better results with a better HSF; I?m currently using a Zalman CNPS7000B-AlCu LED , I?m not to sure since temps at prime stable are 36-37c idle, and 46-48c load are pretty good for 1.475 and 1.5 volts.

yes i quoted myself...


quick update

The bios and mobo monitoring temps were WAY off, i started using Nvidia's monitoring tool (which is good for monitoring and should be unistalled directlty after as the ntune service can effect network latency [at least on xp 64bit].


I decided to back it down to 2.6 GHz at 1.45 volts because I found out @ 2.7 GHz at 1.5v I was idling at 38c similar to before, but at load it was running at 65c OUCH, a 19c difference in reported temps.

Ordered a scythe mine cooler with a 120mm fan instead of the 100mm silent fan that comes with it, as silence is not a worry for me.

Temps now at 2.77ghz prime stable @ 1.475v as per nvidia?s monitoring tool, 36c at idle, and 52c (yesterday with no AC on) load. I?ll have to check again now that the AC is working again in the house, but I?m assuming it will drop 2-3c. almost stable at 2.8 @1.5 v but the second core errors out after a while in prime.

Now I?ll probably back it down to 2.75ish and live with it there, as I like to find the limit and back it off just a bit to ensure for temp spikes and the like.

The scythe mine cooler if prett sweet. And much lighter then the ninja or TT 120. 560 grams as compared to 665 for the ninja and 900+ for the Tuniq Tower 120.

good air cooling round up - summary page

start of the roundup

I <3 my Scythe Mine`.

 

[Steaksauce]

I think my 4200 went up to 2.5 or 6 on stock but running it at 2.74 at 1.45v at 35 idle. I just got it couple days ago so I have more testing to do.

 

[Noubourne]

Just got mine a couple days ago. On stock vcore I can boot in at 2.7Ghz, but it isn't prime stable.

I haven't tested much below 2.6Ghz, but I can tell you it doesn't do 2.6Ghz stable at stock vcore.

At 1.4v it's pretty stable at 2.6Ghz though - it hasn't failed the tests I've run so far.

Replacing my Zalman 6000b-Cu with a Scythe Infinity early next week to see what I can do about my high temps. At 1.4v I hit 55C load with the case side off, and 58C load with it on. It is not stable with the case side on. 14hr primed at 2.6Ghz with it off though - pretty much 55C core temp the whole time.

Asking how high it will go at stock volts isn't quite descriptive enough. Different people consider different things "stable". I can boot at 2.7 at stock volts. Does that really mean anything though? It errors the moment I start SuperPi.

I'm SuperPi stable at 2.6Ghz stock volts. Not Prime stable though. Never had it crash on me with stock volts at 2.6Ghz, but my requirement is 10hrs of dual-Prime (Orthos), plus stable in every game I own, 3dmarks, and at least one round of Memtest. The idea is that I should never have to worry about stability, so I go overkill. There are people out there doing near-suicide screens or just validating CPU-Z at whatever speed before it locks up on them, and calling that an overclock. Those are meaningless imo.