Dang, this 3800+ X2 is sweeeet!

NoxTek

The Geek Redneck
Joined
May 27, 2002
Messages
9,300
Sorry to the mods if I put this in the wrong forum, I wasn't sure wether to put it here or in overclocking and cooling. :)


So I got a 3800+ X2 (939) for a client and have been dicking around with it today trying to see what I can get out of it. It's got a stepping code I havent heard of or seen catalogued yet at any of the overclocking sites:

LDBFE 0610VPMW

Cooling this processor is a Zalman 9500 LED w/ AS5 on the die.
Mobo is an eVGA nForce4 SLI (NF41)

Here are the results sofar at stock voltage:





Granted I've only been stressing for 2 hours, but still... I can't wait to see how far it can really go. :D

I might have to just keep this for myself and figure out a way to pay for it. :p


<EDIT>

The CPU came from Newegg and was shipped overnight to me on Thursday (recieved Friday). It's from a retail kit.

</EDIT>
 
I'm sure someone is going to lambast you about overclocking a client's CPU; it just won't be me! :)

Good overclock on the stock voltage...
 
Haha, uh oh..you're under the X2's spell now. Doesn't look promising for you. ;) Going from a Winnie to that will def. make you rethink things. Looking solid so far!!
 
Yeah I just had to pick up my 3rd 3800x2 offa Ebay (OEM used for $135 delivered... couldn't resist). Although There's a Conroe somewhere down the line waiting for me, the current prices on the X2s make them very attractive. Just a great budget chip, even coupled with a budget ASRock board.
 
I have to say that is a very impressive result from a 3800+. Then again, my friend just got a 3800+ box and I have been surprised by how responsive it is, even at stock. Of course I will have to one-up him with a Conroe box soon, but for now he does have a killer rig he was able to get for only $700.
 
Did you tweak your bios alot to get that out of what looks to be stock voltage? I have a 0610 4200+ and it doesn't really like going over 2400 on stock volts! And even bumped up a smidge i only hit 2450 or so stable. Did you burn in by chance as well? :)
 
Tazman2 said:
Did you tweak your bios alot to get that out of what looks to be stock voltage? I have a 0610 4200+ and it doesn't really like going over 2400 on stock volts! And even bumped up a smidge i only hit 2450 or so stable. Did you burn in by chance as well? :)

No burn in.. the CPU was just recieved from Newegg yesterday, no tweaks in the BIOS aside from the fact that my ram is running a divider. I might post screenies of my BIOS settings later just in case I happen to be doing something wrong, you guys can set me straight. :)


Anyweay sadly after about 6 hours on the stress test in the first post, she locked up... :(


So I decided to play with the voltage and see what I could do. Just for curiosity's sake, I just wanted to see how high I could go and have the rig even boot windows to the desktop. I cranked it all the way up to 2800mhz and then began adding voltage until it booted into Windows, and then added another smidge of voltage just to add for stability:






I don't expect it to stay stable long at this speed, but it still amazes me that it even booted at 2800mhz. As you can see I had to lay on the voltage though, 1.55v. I'm not sure what's exactly safe for these 90nm cores, my old Winnie would hardly break a sweat at 1.55v.

Yes, I am keeping this CPU for myself... I'll buy another for the client build. :D
 
WTF HAX ;)

My bitch of a 4800 just dont like to go very far, it either shutdowns within seconds of boot or wont boot and gives me bluescreen, I figure its teh ram that be doing it to me, either way I am stuck at 2.6.

But I am happy with vid cards, auto detect gives me a rise from 450/1250 to 514/1345 thats stable, and they dont get any hotter except maybe 1 or 2 oC than stock :D
 
My 4600+ (Socket AM2) can only do 2.6 as well. It takes 1.45 vcore to do so and gets too hot to attempt any further. Running two instances of Prime 95 with an ambient of 70 deg F, my Zalman 9500 will reach 140 deg F under load!

I should've grabbed a 3800!
 
2.6 @ 1.375v and 2.8 @ 1.56v??? i really hope your mb overvolts because that it a lot for 2.8

i say give it to the client, thats just a mild OC. you need to find one that will do 2.8 @ 1.5 or a CCBBE opteron that will most likely do 2.8 on stock volts.
 
So I just woke up (yeah, I'm a vampire):





So now I'm just stoked... my question to you guys is: Should I push it further? Add more voltage?

I did a bit of research last night on safe voltages for the 90nm AMD dual core processors and it seems like there are generally 2 camps:

Camp 1: The "Try to keep it below 1.6v as after 1.6v you could dramatically shorten the life of the CPU. After 1.6v the CPU will suffer from electron migration or some such crap." camp.

Camp 2: The "Voltage doesn't really matter so long as you can keep the processor cooled adequately... pump it up!" camp.


Now I generally don't care so long as the CPU will last me a year or so. I personally think that 'premature' death of a CPU due to overclocking is a load of hogwash. I've never had a CPU up and die on me where I would list overclocking as the cause of death, and altghough this is one of my first dual core o/cs, I've been o/cing since the Socket 7 days. :D

So, weigh in guys... what's your opinion?

I also had some asshole pop into another thread last night and inform me that my 3800+ was nothing special at all and most people got theirs to 3.0+ on air at stock voltages. :p


Anyway, long winded post but I wanted to say thanks to everyone chiming in with their encouragement in this thread. :D
 
I'm from camp 1 myself and electron migration is real, it's not crap! Sure it might take sometime for it to rear it's ugly head, but it will.

"IF" you're going to go over 1.6v, i'd switch to a good watercoolling kit or phase change. Of course, JMHO :)

Excellent chip nonetheless :D
 
looks like u can leave the voltage as is and bump the fsb up a bit. i personally wouldnt run my dual core over 1.55v on air even though at 1.55v my load temps are only 45c.

as for that asshole saying most people are getting 3 GHz on stock? not only is he an ass, but he is a stupid ignorant ass too. ive heard of NOONE getting 3 GHz on stock with ANY amd chip. tell him to provide u proof or stfu ;)
 
blazin-asian said:
looks like u can leave the voltage as is and bump the fsb up a bit. i personally wouldnt run my dual core over 1.55v on air even though at 1.55v my load temps are only 45c.

as for that asshole saying most people are getting 3 GHz on stock? not only is he an ass, but he is a stupid ignorant ass too. ive heard of NOONE getting 3 GHz on stock with ANY amd chip. tell him to provide u proof or stfu ;)

At 2.8ghz / 1.55v my temps are as follows:

Core1: 58 - 59C
Core2: 53 - 54C

This is after 18+ hours of stressing

I had no idea that the dual core chips had a temp diode for each core, kinda neat. Anyway I don't think I'll push it anymore until I get a good w/c setup. I've read that anything over 60C is the devil's playground for these chips. :D
 
call me a chicken, but anything over 55c on air is pretty high to me. are you using Everest to measure the individual core temps? 2.8 is a FX-60 already. good job. i guess next task is to try and lower those volts.
 
blazin-asian said:
call me a chicken, but anything over 55c on air is pretty high to me. are you using Everest to measure the individual core temps? 2.8 is a FX-60 already. good job. i guess next task is to try and lower those volts.

I'm using an application called 'CoreTemp" that reads the temps directly from the temp diodes built into the CPU. I've got an eVGA NF4 SLI and while it's a stellar board for overclocking the temperature readings are absolutely crazy. A rather dissapointing bug that left this board in the dust when compared to the DFI.

Yeah, I'm thinking I'll back down on the voltage and probably run this little guy at 2.7ghz on a normal basis. Maybe after a few weeks of 'burn-in' at that speed the temps will stabilize a bit more. Or maybe by then I'll have saved enough for some watercooling. :D
 
So after almost 19 hours priming on both cores at 2809mhz I declared it stable and decided to try backing off on the voltage a tad. Nudged it down one notch to 1.525v and began priming again and within 30 minutes she locked up. Well then...

Now my decision lies in wether to keep it at 2.8ghz / 1.56v or back it don to say 2.7ghz and see how low I can get the voltage. I'll probably make that the plan..... BUT not before I see if I can get it to boot at 3ghz. :D :D :D


Thanks again for all the good comments and suggestions, people! :)
 
well, if u can boot with 3 GHz it would be nice but im sure it wont be close to stable. if you need 1.56 to run 2.8, im assuming about 1.76v MAYBE for 3 GHz stable.
 
blazin-asian said:
well, if u can boot with 3 GHz it would be nice but im sure it wont be close to stable. if you need 1.56 to run 2.8, im assuming about 1.76v MAYBE for 3 GHz stable.


Shit... I better go grab a fire extinguisher!

:D
 
For short term testing likely you're doing on a max out camp 2 is of course fine. But considering this is a build for another longterm I'd camp 1 it and try 2.7 as low as you can go on voltage.
 
GonzoP said:
For short term testing likely you're doing on a max out camp 2 is of course fine. But considering this is a build for another longterm I'd camp 1 it and try 2.7 as low as you can go on voltage.


I decided to keep it for myself. :D

I got another X2 for the customer's build, I'm just loving this too much to let it go. It's a nice cheap upgrade from my old Winnie and it will hold me off until I can either afford an Opteron or go AM2.
 
charliemike said:
What cooling are you running? That's pretty good with vCore that low :)

Indeed, makes me wish I could move to AM2...
 
Personally, I'd run it at stock voltage at 2.6Ghz, since that's a darn decent OC as it is. If you can get 2.7Ghz at stock, that's even better of course, but I'd leave it stock or close. Very nice dude... ;)
 
charliemike said:
I thought he said that was a 939-pin? *shurg*

Not that I can see, and his screenshot with CPU-Z shows an AM2 Windsor core...
 
Don't get greedy. Drop it to 2700-50mhz @ 1.525v and be happy you've got a smoking fast dual core for cheap. :) Pushing the envelope too far is not a good thing - i swear my current chip was stable @ 2.8ghz with RAM at a divider until i tried to boost voltage over 1.6v and run it at 1:1 @ 2850mhz. Never could get that 2.8ghz with a divider back...had to back it off to 2750. I only tried for 5 mins and temps stayed under 60c at all times.
 
The cooling is enough. its a zelman 9500 AM2. As extreme as air cooling gets. The max case temp for this cpu is 70C and it never goes above 50C. Stays at about 35C idol. The voltages are rigth 1.4v even 1.35v is enough just making sure. And I been running it at 1.4v ever since. This chip isn't limited its the bios issues the bios has going over 280HTT. If I wanted to it would run at 279HTT just fine. lol Really strange. But yes its AM2 aright. That windsor core is a FX core for ya. Wonder why it can run at the same speeds as a FX-62 on stock volts. ;)

BTW it can run on 1.1v on stock speeds. :D I think I got a EE ver by mistake.
 
blazin-asian said:
well, if u can boot with 3 GHz it would be nice but im sure it wont be close to stable. if you need 1.56 to run 2.8, im assuming about 1.76v MAYBE for 3 GHz stable.

Isnt that a bit extreme? 1.6v is max you want to use if you want your CPU to last for alteast 2 years. But then again it may not hurt it at all cus thats the max safe volts. Tho those chips are different then my 754 cpu I use to run at 1.65v all the time. My windsors max safe vcore is 1.55v and I never really had to use it.

1.76v may kill it if your not careful
 
Back
Top