E6600 Stability accross multipliers/fsb?

Firstly, I'm a noob, so hello all. Enough reading though and time to make a post!

The setup;
E6600 (Thermaltake Golden Orb cooler)
Asus P5b-E
2Gb Kingston hyperX PC8500
7950GT
Thermaltake Eureka Case
blah blah blah

I've finally managed to get my E6600 up to 3.5Ghz but had to do this by droping the multiplier to 7 and setting the FSB to 500mhz. I've also had to set the vcore to 1.5 which is showing as 1.448 in cpu-z and 1.45 in speedfan. Cpu-Z reports this as 3.5Ghz which it it, but Speedfan and Core Temp are both reporting 4.5Ghz (something to do with the default multiplier being 9 I'm guessing!). Anyway, I've spent literally the whole day trying out various settings. The only other way I could get it to boot was at 3.2Ghz (400x8) and I had the ram set at 1000mhz (so a 5:4 divider) but this worked with vcore 1.4. Now I'm running the ram 1:1 but still at 1000mhz. Current Idle temps are 27C.

My question is why can't I get it to boot at 400*9, 420*8, etc or any other speed? It simply hangs, offers a blue screen of joy, or does absolutely nothing (in which case I have to actually unplug the power cord and wait a few minutes before plugging it back in so I can get back into the BIOS). I've been reading a bit about straps but haven't fully understood it enough to make sense of it. Should I just be happy with 3.5Ghz at a 7x multiplier?

I do have some questions about the memory but I guess I should post them in the appropriate forum.

Thanks in advance.

Cpu-Z reports this as 3.5Ghz which it it, but Speedfan and Core Temp are both reporting 4.5Ghz

Click to expand...

i have the same issue on Abit AW9D-MAX. I updated to newest BIOS it fixed this issue.

try to raise chipset voltage up a bit, around 1.5x/1.6v

I have raised the voltage to 1.55v in the bios which is showing as 1.46-1.5 in Speedfan and as 1.238-1.464v in cpu-z (what's with that large fluctuation?). This allowed me to boot at 400*9 giving me 3.6Ghz which I'm happy about. I'm not so happy about the voltage though. I could boot 500*7 at 1.475 so this is a proportionally large increase. Idle temps are now approx 33C.

I have updated to the newest bios available (as far as I know).

I am not Orthos stable at 3.5Ghz or 3.6Ghz so still a way to go...

@pascalbrown:
What stepping did you have?

Stepping 6, Revision B2, Model F, Family 6. (all taken from CPU-Z).

I have no idea what these things mean but I'm guessing that I don't have a particularly amazing processor. It seems to be a bit of a lucky dip with what you get!

Thanks Nasgul. I have already raised it in the NB, the FSB term. and have disabled all the features you mentioned. I still can't get it stable at 3.6Ghz.

Using a 400*8 (3.2Ghz) I am stable, using 1.375v (in the BIOS). Why would I suddenly need to jump up to 1.6v to use 400*9? It seems a mighty leap. Plus I don't want to run my processor 24/7 at 1.6v (that's a little high isn't it?).

1.6v on air cooling is way too high for 24/7. the max you should go for air cooling is around 1.5v.

Thanks for confirmation! I'm now stable at 420*8 at 1.425v (BIOS). Idle Temps are about 29 which is alright I think. I'm sneaking upwards slowly to see what various FSB's need in terms of voltage to keep them stable.

What I don't currently understand is how the 400*8 is stable at a low voltage, but 400*9 requires such a huge voltage increase to simply boot. Anybody got a technical explanation? I'm a noob on the forums but I do have a basic grasp of maths (degree level).

I'm now stable (showing in my sig). I guess I've settled at this since I dont want to pump more than 1.5V into my processor.

I was running at both 3.5GHz and 3.6GHz but couldn't get it massively stable. In the end I gave up increasing the vcore. I guess for everyday use this will have to do. At 3.6 orthos would run but slwoly things would start crashing, first some programs, then some windows services, then the PC would just restart.

I'd still like more info on my above post regarding fsb/core speed and voltage, specifically the relationship (since it seems to be non linear). There must be a sweet spot?