OC problems on Maximus Formula SE
i'm trying to run my first manual OC - spent a month or two reading most everything here and on the ocforums getting a feel for all of the nuances and now i'm trying..
but every time i think i have it it down the comp proves me wrong...
first, my system specs are
ASUS Maximus Formula SE - no WC running bios rev 0803 (er, was it 2...?)
Intel Q6600 G0 on Tuniq Tower + Scythe 88cfm
2x1GB Ballistix DDR2
EVGA 8800 GTX800 @612/1458/945
Corsiar 520WHX PSU
2x Seagate Barracuda 7200.10 250GB in raid 0 X256 stripe
1x Seagate Barracuda 7200.10 250GB Storage
the temps from one core to another only vary around 7C from each other - 0,1,3 are all about the same, 1 maybe 2C hotter on load and 2 usually about 5C lower than the rest
so i believe my HSF is mounted well and that the -5 may be a very small cry for lapping...
my step for step story:
my first step was to determine what all of my stock voltages were using asus probe and checking with other other progs
they are:
Vcore: 1.16 (1.21 per everest) - set to 1.18 as "stock"
-PLL: 1.54
-NB:1.36 (1.39 per everest)
-SB: 1.07 (1.1 per everest)
-RAM: 2.0
-FSBT: 1.23 (1.3 everest)
-ram ref: 1.00
-FSB term: 1.26
i set all of those in the bios plus set my:
-strap to 400
-the ram static read control to enable
-twister to light
-transaction booster disabled with a relax state of 0
-load line calibration enabled
-CPU GTL 0.63x
-NB GTL 0.67x
-and had initially set all advanced CPU featers to disabled, but after initially finding 3.0 on stock the only thing i believed to be stable and reading some more i went back and enabled a couple of items:
Execute Disable Bit and CPU TM Function for added security
note too, that after making this change (and every change for that matter) i ran OCCT for 30min, or an hour, i dont remember now... too many changes, lol. but i did test it regardless
-from there my very first stab at it was to try 9x334 on stock volts and a 1:1 divider right off the bat just to see if it would boot - it did, and even more it appeared to be stable.
-next, after some testing i ran it at 367 but it wouldn't stay stable... i kept having to bump my vcore up further and further until i got to 1.25 (yeah, that is from 1.18 just 0.3Ghz ago...) this was the point my temps were breaching my comfort level - but it appeared stable! ran Prime for several hours and left F@H SMP running @100% for about 12 hours, but day two i decided that i had needed to do the full 12+hours in prime just to be certain i was stable...
this time prime wouldn't even run for a second... nothing... and the system seemed to start to become unstable almost instantly after that
-since my temps were already at the line i booted back in and dropped it back down to 334 and set Vcore to 1.2 and ran the 12 hour test.. over the week i left it at his setting and stayed stable
-this weekend, i used this as my "starting point" - 9x334 @ 1.2
i started by pushing it up to 340 ran OCCT for a 30minute test, then 345 then 350 then 355 when i got an error in the 30 minute test
tried pushing volts up from here, but again even at 1.23something i was getting errors (they did take longer to hit with each increase though)
but at that point i was getting to my temp threshold again and got tired of testing... so backed off to 350...
ran prime (again) stable overnight (~13hrs) and an hour of OCCT stable
i should note that this morning, at about this point in my OC tour, i stopped all of the ASUS boot progs b/c i was having lots of issue with core temp, speepfan, everest... pretty much anything monitor related that i tried to run.
disabling them seems to have fixed the issues with the progs themselves - but could it be the cuase of my correct problems?
now that i am "stable" and have found the wall...
-i set my strap down to 333
run an hour OCCT - stable
-then, i pushed my mem up to a 5:6 of 421x2
ran OCCT for an hour, and then prime for an hour
stable
-so i dropped my timings to 4-4-4-15 and did the same thing
still, stable
used the comp the whole day again and once the wife wanted to TV back i set OCCT to run an infinate to further check stabilty with no real expectation that i would get an error...
but i came back a couple of hours later to a BSOD calling code 050 (apparently memory issues) and the machine wouldn't post claiming "det mem" as its problem... just pulling the plug for 30 seconds got it back up though
the ASUS bios also told me "failed [OC]"
-dropped my mem back down to 1:1 and relaxed the timings set the strap back to 400
set OCCT again, and an hour later came back to the same memory related BSOD!!!
at this point i increased all voltages by a slight amount:
-Vcore:1.21875
-PLL:1.56
-NB:1.41
-ram:2.1 (remember, it is running at less than 400 at this point.. so really i'm -giving it more than it needs)
-FSB term: 1.3
then ran OCCT and got a crash code 4 after 17minutes
(i read somewhere that this code is a fault on behalf of the program, not the system though. however, it is still a memory related crash, so i'm reluctant to ignore it... what do you guys think?)
now, this very second, it is still blinking code 4
not sure what to do from here... please help!
FYI - i ran memtest86 just one or two weeks ago with no errors (after pulling out 2 of the 4 of my sticks that is... was getting errors in the 4.3GB range on just 4GB of ram - i've read that this is normal for ballistix ram on this mobo...)
maybe the 5:6 could have killed my mem, but i wouldn't think so - the sticks should be capable of doing 2x500@2.2 without much trouble - correct?
so my questions are:
-why the "wall" were voltage increases seem to have little/not enough effect?
-why the instances of sudden instability after running fine for an amount of time?
-are there other voltages i can increase on this board to help my case?
-could it be that the mem that i am using on this board is causing all of these problems?
-could my denial of ASUS apps have had an effect on the situation?
-is my use of higher than needed voltages on the ram a cause for any concern?
-are there any other setting that i changed over this trip that may have caused this sudden instability?
i just need help narrowing down where to look really...
