Testing my 478 P4e 3.4@4ghz...

[GRAFiZ]

Well, instead of moving to an AMD system I decided to play around with some of the newer Prescotts. I got a hell of a deal on a 3.4ghz D0 and jumped at it.

Specs:

Abit IC7-MAX3
P4e 3.4@4ghz 1.425v - Air cooled: SP-94 with Vantec Tornado at 3000rpm
4x256 OCZ PC4000 Gold Rev 2 (236 1:1 2.5/4/4/8)
BFG 6800 Ultra OC 450/1180
74gig Raptor
520watt Vantec PSU

Pics:

SuperPi 1meg SSE3 patched... yeah yeah, I know, it's patched. Everyone will use it when the new AMD64's have SSE3.

http://home.comcast.net/~grafiz/superpi.JPG

SiSoft:

http://home.comcast.net/~grafiz/sisoft.JPG
http://home.comcast.net/~grafiz/sisoft2.JPG

I really need to do something about the memory scores, I'll adjust timings later, just got this setup yesterday.

3DMark score:

3DMark 2005: 6189 (http://service.futuremark.com/compare?3dm05=450363)

3DMark 2003: 14282 (http://service.futuremark.com/compare?2k3=3551367)

3DMark 2001: 24575 (http://service.futuremark.com/compare?2k1=8348235)

Anyhow, might not seem like much compared to all the new 5ghz+ overclocks, or the massive 2.8ghz AMD64's, but I think it's still pretty darn fast!

Thanks for looking!

[caLume]

damn fast for 4ghz.
the rams should go better ?

or doesnt the max vdimm push the rams to their limit? / better timings ?

[GRAFiZ]

damn fast for 4ghz.
the rams should go better ?

or doesnt the max vdimm push the rams to their limit? / better timings ?

Ahh... this memory is a pain. The fact is the MAX3 is supposed to push up to 3.2v, but, it's well known it has some sever voltage problems. So, anything over 2.8v VDIMM creates a drag in the VTT, somewhere around 33% instead of 50%.

My board has the MC64 cap removed so voltage is stable, but the VTT doesnt get high enough to run stable.

Basically, I have 2.6v to 2.8v and then 3.2v open for me to use, but 3.2v 24/7 would kill my ram, so I am stuck.

Funny thing is, no matter what setting I use, 2.6 - 2.8v, 2.5/4/4/8 is the best I can seem to do.

It doesnt seem to make much difference on the scores anyhow, timings that is. FSB makes to only big difference. I can also use the 14x multiplier with this chip, but then I have to use 5:4 to get my ram to run stable, so I lose all that I gain.

Up against a wall really, even with 2 sticks removed I dont gain much. I will continue to try though.

:toast:

[cpulloverclock]

nice job but I get the same score than u with a Superpi unpatched

http://membres.lycos.fr/cpulloverclock79/Informatique/4026.gif


and patched

http://membres.lycos.fr/cpulloverclock79/Informatique/29s.gif

[GRAFiZ]

nice job but I get the same score than u with a Superpi unpatched



VERY nice scores!!!

Wow, your ram is set tight, that must be the reason for the difference. Impressive.

[mr_knowitall15]

Graf, do u have the ability to change the multiplier? i have such a similar system, and with the latest BIOS, it "tricks" the cpu into unlocking...kinda. i can select 14 or 17. with the 14 multiplier i have maxed my geil pc4000 out at 265fsb, with 2.9 volts. 2.9 seems to work fine, but i wouldnt try 3.0

[enzoR]

can u not go higher than 4ghz?

[Foxie3a]

Which processors have the option of different multipliers in that board? Only the 3.4E? I have that board, and I still have a 3.0C, but I wouldn't mind playing with a prescott if it let me change the multiplier. Thanks.