P6N Diamond+Quad core overclocking - Help if you have this board.

Question

Am having some issues trying to get even the simplest overclock on this board plus some wierd stuff it is doing with ram settings.  Can't get linked setting to work.
I've read all of the 680i tutorials.  Nvidia, XS, HardOCP..............I've read Toms Hardware pre-release review.  Basically anything I could find..............
Chip is running cool and rock stable at stock settings with either bios.

Answer 1

ok lads - here we go...
it's been a very interesting and revealing morning for me - spent the last 5 hours tweaking the bios, then testing, then tweaking, then testing.  these are the results:
note 1:  unless otherwise stated, all settings are 'optimised defaults' with 'CE1' disabled (go into 'advanced BIOS features/cpu feature' and press F4 if you can't find it) and 'spread spectrum' disabled.
note 2:  unless otherwise stated, 'system clock mode' is set to manual and the memory is set to 800 and timings are on auto.
note 3:  default voltages are used unless stated
note 4:  each prime95 test is carried out for 5 minutes only
note 5:  i'm using a thermaltake kandalf vd4000 case with built-in liquid cooling system.  my water-cooled cpu temp in the bios is 19oC (can't seem to get the degree sign using the tip at the top of this page)
note 6:  cpu is q6600. memory is 2x1gb xms2 dominator pc2-8500.  vga is gainward 8800gtx. psu is hiper 5th gen type-r 770w.  bios version is 1.2.
note 7:  core temp 0.95 was used for establishing cpu temperature. (apparently, this gives default readings that are 15oC higher than the actual relevant cpu temperature).
test 1 (stock cpu setting):
fsb set to 1066
=2.4ghz
error-free prime95 for 5 minutes: YES
max load temp: 46
test 2:
fsb set to 1166
=2.62ghz
error-free prime95 for 5 minutes: YES
max load temp:  47
test 3:
fsb set to 1206
=2.71ghz
error-free for 5 minutes:  NO (4 cores failed within 30 seconds)
max load temp: 47
decided to try pumping some more volts into the cpu to see if i could get it stable in prime95
test 4:
fsb set to 1206
=2.71ghz
cpu voltage increased by 0.0125
error-free for 5 minutes:  NO (4 cores failed within 30 seconds)
max load temp: 47
test 5:
as above, but cpu voltage increased by 0.0250
error-free for 5 minutes:  NO (4 cores failed within 30 seconds)
max load temp: 47
test 6:
as above, but cpu voltage increased by 0.0375
error-free for 5 minutes:  NO (4 cores failed within 30 seconds)
max load temp: 48
test 7:
as above, but cpu voltage increased by 0.0500
error-free for 5 minutes:  NO (4 cores failed within 30 seconds)
max load temp: 48
test 8:
as above, but cpu voltage increased by 0.0625
error-free for 5 minutes:  NO (4 cores failed within 30 seconds)
max load temp: 49
test 9:
as above, but cpu voltage increased by 0.0750
error-free for 5 minutes:  NO (4 cores failed within 30 seconds)
max load temp: 49
test 10:
as above, but cpu voltage increased by 0.0875
error-free for 5 minutes:  NO (4 cores failed within 30 seconds)
max load temp: 49
test 11:
as above, but cpu voltage increased by 0.1000
error-free for 5 minutes:  NO (4 cores failed within 30 seconds)
max load temp: 50
this took me to the recommended limit for voltage for the q6600, which is 1.375v.
it seems i can safely say that changing the cpu voltage makes absolutely no difference at all.
still keeping the fsb at 1206, the next thing i tried was increasing the NB voltage to 1.275v and keeping the cpu voltage at +0.1000
these settings still couldn't get prime95 to run error-free for more than a few seconds.
keeping the fsb at 1206, cpu voltage at +0.1000 and the NB voltage at 1.275, the next thing for me to try was increasing the 'fsb vtt voltage' by 2%
SUCCESS!!! SUCCESS!! SUCCESS!!
using the above settings, prime95 ran error-free for 5 minutes.
my next problem was trying to work out which of the above voltages settings was the kicker.
keeping the fsb at 1206, i returned the cpu voltage back to default and kept the nb voltage at 1.275 and the 'fsb vtt voltage' at 2%.
SUCCESS!! SUCCESS!! SUCCESS!!
prime95 ran error-free for 5 minutes.
next test was keeping the fsb at 1206, the cpu voltage at default.  i also returned the nb voltage to default but left the 'fsb vtt voltage' at 2%.
SUCCESS!! SUCCESS!! SUCCESS!!
prime95 error-free for 5 minutes again!
it seems like the vtt voltage is an important player in getting the system to appear stable.
the next test, i set fsb to 1254 giving 2.82ghz.
keeping all voltages on default except the 'fsb vtt voltage' which i kept at 2%.
prime95 produced errors again, pretty much within the first few seconds...
the next test, i kept fsb at 1254, with default voltages, but the 'fsb vtt voltage' was increased to 4%
SUCCESS!! SUCCESS!! SUCCESS!!
prime95 ran error-free for 5 minutes.
all of a sudden, it seems that regardless of the other settings, the only voltage that matters is the 'fsb vtt voltage'
next test, fsb was set to 1278=2.87ghz.
vtt voltage was 4%
prime95 ran error-free for 5 minutes: YES
cpu temp during load: 49o
next test, fsb was set to 1302=2.92ghz
vtt voltage was 4%
prime95 ran error-free for 5 minutes: YES
cpu temp: 49o
next test, fsb was set to 1326=2.98ghz
vtt voltage was 4%
prime95 ran error-free for 5 minutes: YES
cpu temp: 49o
next test, fsb was set to 1350=3.03ghz
vtt voltage was 4%
CPU FAILED TO POST AND REQUIRED BIOS RESET
next test, fsb was kept at 1350=3.03ghz
vtt voltage increased to 6%
booted fine, but computer froze during prime95
next test, fsb was kept at 1350
vtt voltage was increased to 8%
again, pc froze during prime95
next test, fsb was kept at 1350
vtt voltage was increased to 12%
computer failed to post, requiring reset.
12% was as far as i was comfortable going to, as i'm not sure of the consequences of having this set too high.
it seemed i had reached the limit as far as pushing the fsb goes.
i returned the vtt voltage back to 4% and lowered the fsb to 1334 (i had a hunch that anything over 1333 would give me problems).  again, this failed to post.
i set the vtt voltage to 6% - again, it failed to post.
i set the vtt voltage to 8% and again, it failed to post.
i lowered the fsb JUST ONE NOTCH to 1333 and set the vtt voltage to 4%.
this gives an exact cpu speed of 3.000ghz
i'm not sure if this is a coincidence or not, but this setting posted, booted and ran prime95 for 5 minutes without any errors.
it then ran prime95 for over an hour with no errors.
curiously, i set the fsb back up a notch to 1334 and it refused to post, requiring a bios reset.
as i said, it seems 1333 is the absolute limit that i can set this board to in my system configuration.  now i don't know about you, but the fact that 1333 works perfectly stable, and 1334 doesn't even post tells me that there is something not quite right with this board.
is it too much of a coincidence that the cpu/mobo hits a wall at EXACTLY 1333/333/3.0ghz?
the next thing i did was set my memory to 1066 and manually set the timings to 5,5,5,15.
coupled with my new rock solid 3.0ghz q6600, this is about as far as it is physically possible for me to push this setup and i'm more than happy with the results.
now i can't wait for the G0 stepping to see what that can do.
maybe a future bios will help with this - who knows...