Overclocking Results: Super-PI Challenge! (please read the first post)

Running at 166 FSB might have something to do with it

The RAM can run 200 at stock - tried that?

 

I know, I have tried may differant multi's, FSB's & DDR FSB's, Look here It doesn't seem to make a differance T"B"H

 

how much do i have to worry about additional heat with the increase of voltage on my RAM, because if I can get that higher than I can probably go somewhere around 2.5GHz
maybe get my time in the 35s or 34s

 

your ram should be fine with avoltage upto about ~2.8-2.95 just make sures there's decent airflow

 

ok sounds good I will up the voltage once i get back to my house where the temps are much colder
what about increasing the voltage for my CPU?

 

Well that depends on your CPU cooling, I'd say 1.4~1.5V on stock air cooling, 1.5~1.65(MAX) on good aircooling(like a xp-1200, zalam 7700cu.etc) or watercooling

EDIT: Please look at this thread if you havent already !

 

thanks for the link I will be sure to up the core volt some

 

On my step-brothers Quad-Core 2.5GHz G5:



26.961s

 

upped the memory volt and the core volt and got it around 2.55GHz, best time was 34.656s
the HTT won't go any farther anyone know what I can do to get it higher?

 

My opteron 146 pulls a 33, but I'm not on it right now so I don;'t have a screenshot.

 

have you lowered the HTT multiplier? the defualt is x5 so when you run @ 200FSB, its 200*5=1000, then you may up the fsb to ~300 thats 300*5= 1500= thus you lower the htt multi to x3 the is 300*3=900 . in other words keep the HTT total under 1000 . that could well be the problem but correct if i'm wrong.

 

thx I will try that but I am currently running prime 95 torture test to test an overclock

 

nice score, but sadly not comparable as its a different version

 

new score

specs as per sig

 

Could you find me a correct version that will run on OS/X?

I got the version I used from here.

 

he means that, because there is not an identical copy of superpi x86 for powerpc's, its impossible to compare as it's code may vary and thus will produce faster/slower results.

new run:


That's about tops it will do on this ram, and on air cooling at that (haha gotta slap a delta on my XP-120, only got a low speed panaflo atm). Ram won't go a single mhz more at those timings. I have to set it to 3-4-4 (impacting the speed) to get more out of it, and then it only goes to about 280mhz. dropping to the 183 divider lets me max out the CPU at ~3040mhz @ 1.6v, but then again its still slower because of the slower ram.
Still happy with it . Pretty good considering i'm less then 1.5 seconds off pookeyhead who has pretty much the best hardware money will buy atm, watercooled aswell.

 

right, I have been thinking to myself over the past few weeks that I should really update the table, take out the old times and put in the times with the new superpi etc etc.

I just can't seem to find a good chunk of time to do it, how I miss the endless spare time that is university

This means its delegation time...

If anyone would like to take over as the official Bit-Tech keeper of all things superpi, drop me a PM

 

Do ram timings have a big effect on PI? My 146 at 3ghz id doing 30s with 2GB of ram at 3-4-4-8 I see TMM is beating that with his CPU clocked less

 

Firstly, I'll be working on the new result table tonight, it will be the top 40 in this thread

Secondly TMM has his memory running @ 272Mhz, It's ushually the higher clocked your ram is the better score but timings help abit.

Finally, heres my latest:

-34s
please note Vcore is actually 1.6v NOT 1.8v

 

33 second Venice 3000+ run, near-stock volts