Does this seem OK?

20 Mins at 300FSB, everything else still the same and temps are all around 50c.

I know I have the step thingy on (so when not pushed CPU-Z shows 300 x 6) but I am wondering why Core Temp shows 450 x 6?

I am trying to keep memory at 800 with dividers for the moment, so I guess just increase FSB until Prime fails and then up the Core Volts slightly and try again? Do I need to worry about any other BIOS settings for the moment?

Cheers
 
OK,

334FSB is stable at core volts at temps 54/54/50/54 (why is core 2 always lower?)

If I jump to 3.2Ghz I start to run into problems, there are a variety of Volts to adjust in bios, so I upped CPU (and RAM) by the min margin, but still get errors as soon as prime starts.

I don't really know which volts to what, and what ones I should fiddle with.
 
Q6600 can have upto 10c difference in coretemps on load... this is normal... can be to do with how even the heatspreader on the CPU is or how well you heatsink sits on the CPU.. tbh 4c on one core is pretty good and nothing to worry about


ok... volts :)

vcore = CPU volts normally a prime fail is because the core under load @ the higher clock doesnt have enough power to function correctly.. raising your vcore will increase your load temps as more volts = more heat.. is best to raise it by the smallest amount your motherboard allows and try the test again keeping an eye on the temps

RAM volts (vdimm on my board) should always be set at the manufactures recommend level you shouldnt need to change this at all!! RAM is far more likely to be damage through overvolting than your CPU

Northbridge volts, northbridge controls the link between the CPU and RAM, should you be failing on prime95 through rounding errors is more than likely this is a RAM and not a CPU issue, your first step should be to up the NB volts slightly.. if it still fails then slacken the timmings on your RAM
 
Currently Ram is 5-5-5-18 at 2.1V
Bios Readings
CPU 1.3250V
CPU VTT 1.20V
MCH 1.25V
ICH 1.05V
ICHIO 1.50V

So I will up the CPU Volts a tad, turn of rounding errors in Prime and try again..
 
ok well shouldnt effect your clock to 3.2ghz if you keep getting rounding errors dispite vcore increase then might need to up your NB volts a little
 
I am at 1.3350 Volts and 3.1Ghz, Prime runs - but then has an error - not a BSOD, but errors out.

How far up should I push vcore before I up the NB Volts (is that the MCH one?)

Does it matter that my Ram is o/c slightly at this FSB.

Thanks for the help btw :D
 
depends on the RAM could be that it is the RAM failing would lossen the timmings a little... not sure what the NB voltage is on that board maybe someone else can tell you... but lossening the timmings will have the same effect and you can fiddle with NB voltage later... if lossening the timmings doesnt work then raise the vcore.. recommended limit on air is 1.55... but you will max out on load temps way before then I would guess
 
What's the method to lossen timings on RAM, they are 5-5-5-18 now, do I go 6-6-6-18 or 5-5-5-20?

I could use a 1:1 Ratio, which would drop them down to 670Mhz or so...
 
Using 1:1 allowed me to pass 3.1Ghz with 56/56/52/56 temps under load.

Going to try 3.2 as the closer to 400FSB the closer I am to 1:1 Ram being a decent speed.
 
Gone back to 3Ghz for the time being with everything at normal volts (4-4-4-12) timing. Getting scared that after 3.1Ghz was getting up to 1.4050V trying for 3.2 - so thought will let things settle for a day :)
 
volts seem a little on the high side... can you post your BIOS vcore, and then from CPUZ a reading of the vcore @ idle and then @ prime95 load..
 
ok...

This is at 3Ghz so 334 X 9 (1:1.20 Ram @ 801)
CPU From Bios 1.3250V

Core Volts From CPU-Z @ idle = 1.150V
Core Volts From CPU-Z @ load = varies around 1.260V - 1.280V (occasional peaks at 1.290V)
 
Not at this FSB as that I think is the min I can drop the core to..

When I up the the FSB to 356 I was using 1.4050V in bios, was nervous to push it up more as it seemed a big jump and still was failing prime.

346FSB was stable (at least for 10 mins) at 1.3350V so maybe that's not such a huge jump and I should up the core from 1.4505?
 
You must log in or register to reply here.
Back
Top Bottom