Q6600 vs Q6600 with some overclocking.

Get physical with some of these projects? Asteroids, Cosmology, Einstein, LHC, MilkyWay, Test4Theory,
Post Reply
jaysback
Posts: 202
Joined: Thu Mar 16, 2006 12:00 am

Q6600 vs Q6600 with some overclocking.

Post by jaysback »

Found the workunit had run against another Q6600 at standard clocks (2.4) vs mine @ 3.4
At least on einstein credit is pretty stable, with seti I never get granted my claimed credit.
MikeMarsUK
BOINCSynergy team member
Posts: 110
Joined: Sat Aug 12, 2006 1:00 am

Post by MikeMarsUK »

Prime95's torture test is a very good way of testing stability.  I use a 24-hour run (one instance per core) to test whether the overclocking is stable enough to run the climate projects.  

I get 1.8kcredit/day with the SAP (HadAM3) model, or 3kcredit/day with the CPDN/HadSM3 (slab) model, which runs particularly well on Linux and the new Intel architecture.  Pity that I'm not keen on slab and hence only run SAP on it :-)


What is your setup like?  I haven't got the Q6600 stable over 3GHz yet (Can boot a lot higher, but WUs / torture test always errors out).  Mine is just aircooling (Ninja B+ with a high flow 120cm fan strapped onto it). Doesn't help that the box only has Linux, and I haven't figured out how to get temperatures yet, lm_sensors doesn't seem to give me any numbers.
jaysback
Posts: 202
Joined: Thu Mar 16, 2006 12:00 am

Post by jaysback »

I used to use prime but now i just go with if it runs boinc o.k. doesn't error out of w/u and i can run games while crunching then i am a happy camper. lol.
Just been playing with it all morning and i think the bios is getting fed up with me changing settings all the time. lol.
Currently 8 x 370 2959 @ 1.325V set in bios but Vdrop takes down too 1.24. Or at least thats what asus probe is wanting to tell me.
Temp wise cpu 49 m/b 32 under load, ambient temp 27.

Can't get it stable now back at 3.4
Was having a play cause i am out for the weekend and wanted to leave it crunching and not cooking.

That will teach me........
Post Reply