Athlon 64 3400+ overclock

[joe2004]

I got one coming shortly but I would like to know what kind of overclock I can expect. I am hoping for at least 2.5 GHz for that is the reason I bought it in the first place.

 

[Treyshadow]

Good luck, I have not overclocked mine, neither have any of my lan mates (we all got the same setup 3400+ and MSI FISR2)

Well real overclocks. Mine is running at 2310 right now. But I wasn't trying to OC. I have it on a silent heatsink., a Zalman 7000 Cu, but some of the guys have them on Silent Boosts. All with about the same temps.

 

[Shimmishim]

if you're air cooling, 2.5 ghz might be possible... but you may need to up the vcore...

the highest i've seen on air is around 2.45 so... 2.5 is VERY possible!

 

[joe2004]

the highest i've seen on air is around 2.45 so... 2.5 is VERY possible!

It is on the way, tomorrow I'll get it. If you check overclockers.com database there are two guys there getting 2.5 GHz on air. I have Asus K8V and the voltage available is up to 1.75V. By the way I would not go any higher. My air cooling solution is TT Venus with Tornado which is as good as it gets I think. So anything less than 2.5 GHz would be a disappointment to me. I'll report shortly.
I need to get Prommie like you, probably this summer.

 

[Dman877]

No agp/pci lock on that board so your cpu may not be the limiting factor.

 

[joe2004]

No agp/pci lock on that board so your cpu may not be the limiting factor.

Nope. I can run 248 FSB (there isn't really FSB). Since 3400+ has multiplier 11, that would be 2728 MHz. No Athlon 64 can do that, short of phase change cooling. So I'll be OK if only CPU can take me to 2.5 GHz.

 

[joe2004]

Testing it now. I think AMD 64 has major problem with memory controller that isn't capable of running many banks at the high speed. I know this already. AGP/PCI lock is not as big problem as some people think. I am getting it at 216x11 = 2.376 GHz which matches the performance of Athlon 64 3200+ at 242x10=2.42 GHz due to the fact that I can run memory synchronuously.

Finally at 223x11=2.453 GHz 3400+ is beating 3200+ SuperPI performance. I get 84 seconds for 2 MB which is my previous record with 3200+ running at 2.48 GHz.

Getting 225x11=2.475 GHz, almost there, running SuperPI 2MB for 83 seconds, new shinny record for me. The same at 226x11=2.486 GHz. Sandra 2004 CPU integer ALU is 10656, nice 2% gain over 3200+ at 2.42 GHz.

 

[Shimmishim]

also think of it this way though....


was $100+ worth it for 1 second in superpi?

and i can run 242 mhz fsb 1:1 synchroneously...

maybe you need a vdimm mod to your board

 

[joe2004]

was $100+ worth it for 1 second in superpi?

Hey man, of course yes. I am a fanatic you know. But the fact is that 2.48 I could run with 3200+ with only one memory bank which kills my hash size in chess. Now I am running two banks with 512 MB which is minimum I need.

Now getting 227x11=2.497 GHz, Sandra ALU does 10,736, and this is on air cooling!

But here is the bad news. Prime95 stability is only at 224x11=2.464 GHz which is very small gain over 3200+ that I could run Prime95 stable at 2.42 GHz with two memory banks. Using 1.66-1.68V core if I can believe monitoring..