processor glass ceiling?
i have my 2.8 at 3.23 right now, and i'm prime95'ing it to double check it's integrity. i tried nudging it 3.26 and it booted, but didn't test it. then i got it to 3.3 and windows just took a crap on me right when starting up. so, i turned the ram to a 2/3 ratio (in my bios it's listed as 1.66, which i took to mean 2/3) still no good. i'm just wondering if i can get it any higher, because this is most definitely not a temp problem ... even at 3.2 i idle at about 22*C. as of the minute i'm writing this, my fsb is set it 232. that puts my bus at 927.8 (i think the strange number may be due to MIB:memory intelligent booster...should i turn that off?). my ram timings are set to By SPD. now, i just realized i could play with voltages, but that makes me incredibly nervous, and i wouldn't know if i should up the ram or the cpu voltage. i think i'm stable a 3.23 right now cuz prime hasn't found an error in the last 20 mins or so. so, upping my voltages on the cpu will raise the temp by how much approximately? so, a summary:
1)can i go any higher?
2)if so, how?
3)kill MIB?
4)how much, per increment, will my temps go up and what effect does upping the voltage have on the cpu?
thnx all
1)can i go any higher?
2)if so, how?
3)kill MIB?
4)how much, per increment, will my temps go up and what effect does upping the voltage have on the cpu?
thnx all
0
Comments
if you're stable at 3.23 with your cpu then you can probably squeeze some more mhz out of it by upping the vcore, i'd think (assuming you're using a good HSF) that 1.6v vcore - AT LEAST - is completely fine. however, when you hit a ceiling where you need big vcore raises for small speed raises, THEN you've hit your ceiling
OK, some questions?
What is your CPU temp at your achieved overclock?
What speed is your DRAM? PC3200? Manufacturer? Brand?
What power supply unit do you have? Probably not a limiter if you're running OK at 3.23GHz.)
Boosting voltage on the DRAM won't hurt it at all. The default memory voltage adjustment increments built into the board/BIOS are not large enough to cause damage.
Try boosting both DRAM and CPU voltages two notches each. If you achieve a higher, stable overclock, you can always back off the voltage and test for stability. (Please do though, report your temperatures before boosting CPU voltage.)
OCs on P4 Northwood 2.8s average around 3.2, but 3.4 is not all that uncommon. I can get an almost stable 3.6 out of mine, but I've got a very well-ventilated case and an above average 2.8.
BTW, stable overclocked Northwoods in hyperthreading mode are monster Folders.
Self-test 8K passed!
FATAL ERROR: Rounding was 0.5, expected less than 0.4
so i think i'll knock up the voltage, watch the temps and try again...but bumping it doesn't have any adverse effects, does it? (it won't kill my cpu more than just regular overclocking?)
pc3200 ddr Corsair Value Select ram, not sure what brand the chips themselves are but i can check.
power supply is a 500W Ultra.
i think i'll up the ram and cpu one voltage notch each and try again (will post temps as well) if that doesn't work, i'll go one more and try again. but i'm not sure how far i want to mess with that.
\\EDIT: when running p95, i selected blend. i did get an error once in memtest86, so after i let prime run for a bit on 3.2 ghz (i declocked it some) i'll try selecting the first option, where very little ram is tested and see what happens.
What, are you high? Maybe 100*F, not C. He says room temp is 22.5C. I don't know what that equates to in *F, which is what I go by here in the US.
I think we're both in the same boat entr0py. 2.8's stuck at 3.2 until someone gets a new HSF. (I've been considering the sp-94.) Not that that's a bad thing! 400 free MHz is really cool!
:usflag:
Only at those settings will you figure out how far your CPU can actually go (well, at least on your current cooling setup) without your memory holding you back.