strange prescott d0 behavior..

[mahiman]

ok, just got a 3.2 d0 (about 2 days ago)....

i have an ic7g (max2) (bios 25)
2x512 mushkin pc4000 (rev.a) (2.8volts)
enermax 550 watt power supply
waterchill/l30 pump (120mm rad sunon fan)

here is the funny part...
the d0 will run at 3.6 stock (which in the bios says its around 1.3-1.31)
BUT..the minute i raise the voltage the processor runs slower...sisoft benches slower than a 2400+....but at stock voltage the scores are fine...

if i bump the voltage up, 1.53 or higher it would cut off the minute i hit the wondows boot screen...it posts at 4.0ghz...but blanks out after the dmi screen...(update success)...

sadly i figured i would be able to make it to 250 1:1 on this thing but maybe i am doing something wrong...or maybe it hasnt "broken in" yet..or is this some wierd behavior of prescotts i have yet to hear about...not only does sandra bench low (we know sandra hates prescotts anyway)...games play slow..slower than my 2.4c at 3.3...tried dividers everything..maybe just not the right combination...or maybe a little while of pumping some electricity through her will soften up this wacky behavior...

i lurked here for months and decided to go with the d0...but they dont seem soo good after all...or maybe one of you intel experts can chime in and help me out..cause i am definately confused..

ps: it seems like when i bump the voltage the performance goes down..and its not hot...at 1.53 volts or above it will just reboot...any advice or experience welcome..maybe we found something new here...

 

[Fox3a]

Voltage dosen't always make things better. If its not stable at a higher voltage, then keep it down. Cooling might not have anything to do with it.

 

[Nitro16]

If your CPU is benching at lower than a 2400+ when you raise the voltage it is most likely that it is getting too hot and throttling, however considering your cooling is relatively good its very strange that this is happening.

You also mention that your 2.4C @ 3.3 was faster. This is actually quite understandable, the 2.4C is a Northwood (slightly faster than Prescott clock for clock) and also running at a much higher FSB than your 3.0E @ 3.6. Sadly, due to the extra long pipeline of the Prescott, games are definitely not a strong point of this CPU.

Its a shame that you havent seen better results, many were looking forward to the new stepping, however much like the D1/m0 Northwoods, stepping on Intel chips doesnt seem to have much bearing on how well they overclock.

 

[unhappy_mage]

it may not be "burned in" as you say. try running folding on it 24/7 for a week and then try again. then let folding keep running

iirc sandra also has a "burn in wizard" which does basically the same thing. however, sandra isn't real good at benching anything, so i'd let it go asap.

 

[mahiman]

funny enough, its getting higher..got 3.8 stable now...i am figuring out the little pattern of this thing...and at 3.8 this baby screams....at 4.0 (which i made it to but had to up the vcore..maybe some more burning in (thermal paste settles..etc)...i think these things shine after 3.8 and intels engineers knew this...i was getting worried..but you guys have given me some insight...after i get her stable ill post benches at 4.0..if she doesnt make it there stable..forget i exist.. (lol)...thanks for all your help..now i need to stop eating and go to sleep (too much coffee last night...


ps..btw...the closer you get the prescott to 70 degrees the slower it runs...further you get it away from 65 it passes my buddies 2.8c@~3.5 in fpu/alu (she is at 3.6 at that time, stock vcore)...but maybe these prescott benches slambing it arent taking this into account...ill put it this way...i am typing this fully while running prime (only 1 instance) not bad...couldnt do that with my northwood (though 500mhz slower...but leaps and bounds faster in the feeling)