Max core voltage for AMD Athlon 64 3200+ Venice?

kaisellgren · Jan 17, 2008, 09:16 AM

Hello,

I am overclocking my CPU, therefore, I need to know the maximum vcore I can safely use. AMD's official website sucks and I am unable to find the information. My CPU is AMD Athlon 64 3200+ (default 2000 MHz,200x10) and it has 512 kb cache, E6 stepping and uses 92nm technology.

I am using vcore of 1.575 at the moment and my CPU is not stable at 2.40 GHz (240x10). My goal is 2.40 GHz right now.

Thanks for your help!

Zelig · Jan 17, 2008, 11:25 AM

Anything past stock voltage voids the warranty.
I would in no situation advise going past 1.6, asuming your temps are good.

Personally, I probably wouldn't try past 1.5, and wouldn't run past 1.4 for 24/7 usage.

Generally when I'm overclocking processors, I keep voltage at stock, and just take the frequency as far as it will go from there.

edit: Those cpus generally reach a point where small MHz gains require greater and greater voltage increases to get stable, it generally isn't worth upping the voltage by 0.3v for an extra 200 MHz.

**Grace** · Jan 17, 2008, 11:25 AM

Hello Kai,

There is no such thing as a maximum safe voltage. The only safe voltage is the stock voltage. All voltages higher than that will reduce the lifetime of your CPU...supposedly. A CPU will last for many years even at voltages much higher than stock when the cooling is adequate.

In short, I'd try voltages up to 1.7V if I were you but watch your temps. Over 65C degrees at full load are not desirable.

PangingJr · Jan 18, 2008, 03:33 AM

Quote:

Originally Posted by kaisellgren

Hello,

I am overclocking my CPU, therefore, I need to know the maximum vcore I can safely use. AMD's official website sucks and I am unable to find the information. My CPU is AMD Athlon 64 3200+ (default 2000 MHz,200x10) and it has 512 kb cache, E6 stepping and uses 92nm technology.

I am using vcore of 1.575 at the moment and my CPU is not stable at 2.40 GHz (240x10). My goal is 2.40 GHz right now.

Thanks for your help!

you could use a processor stable and stress test program likes StressPrime 2004 Orthos or Prime95 to find the minimum necessary core-voltage requirement for using in the particular processor overclocking and at a certain processor overclocking speed, and then you could add just a little bit more volt to the required core voltage.
however, while the processor is overclocked, like suggested, keep checking the processor core temp, and try to keep the core temps at as low as possible and you should be okay.

[hobo]eclipse · Jan 18, 2008, 04:07 AM

Since your runnin at 240*10, what the situation with the RAM? is it 1-1 or is divider in use? did you lower the HT to 4X? could be a chipset voltage issue. BTW isnt the stock volt on a Venice 1.4 or 1.425. Just seems like alot of voltage for a 400MHz OC, Im betting you can lower the CPU voltage and tweak the RAM and chipset to get stable. Could be wrong though! LOL

kaisellgren · Jan 18, 2008, 06:24 AM

Hi,

Thanks for your replies. I have two pairs of PC 3200 400 MHz DDR ram in that PC and I think I am using 1:1 ratio.

I am using Zalman CPU cooler called CNPS7000AlCu. The idle temp is around 32-34C and the load temp is something like +10-15C more. My PC did POST and boot and I launched 3DMark 05 and it was working fine until it started the CPU test.

I saw actually no change between vcore 1.45 and 1.575 so I think the vcore is not the answer in here. I'll try to tighten or relax the RAM timings and see if they get it more stable. Unfortunately, I have no clues, which or whose memory sticks are those. I bought those two memory sticks of PC 3200 from a friend and I do not know the manufacturer.

I saw a couple of posts here and there where they told they were able to reach 2.8 and some even 2.9 GHz on 3200+ Venice and with a aftermarket cooler.

Also, do you think it could be an FSB hole? I had an FSB hole in my P5N-E and Q6600 so I was just wondering...

Zelig · Jan 18, 2008, 07:05 AM

Quote:

Originally Posted by kaisellgren

I saw a couple of posts here and there where they told they were able to reach 2.8 and some even 2.9 GHz on 3200+ Venice and with a aftermarket cooler.

Not too likely, those are rare cases.

Don't be afraid of dropping the ram multiplier, internally, it's running on a divider from the total CPU speed, so there's no "real" multiplier. As such, there's no performance difference between running 1:1 and otherwise, if the ram is at the same frequency.

[hobo]eclipse · Jan 18, 2008, 07:18 AM

yas yer best bet with the unmatched RAM for stability would not be to try to OC them but set the divider back to 166/333DDR, that way yur RAM will default to somewhere around 166 but the processor (the part where you overclock) will still be 200*X. you up the CPU bus (200) and the RAM follows but starts from 166.