I know there are a lot of differences between chips even of the same type depending on batch, but I'm getting really frustrated. Other people are getting much higher clocks and I can NOT get mine stable. I first tried at 3.6 and voltage up to 1.51 still wasn't stable, and I didn't think it wise to go beyond that. Then, at 3.4... still wasn't stable all the way to 1.51. Now I'm working through 3.2 ghz and I'm at vcore 1.4 and it still isn't stable. I've pumped my FSB up to 450 and put my multiplier at x6 and everything appeared stable (15 minutes on Orthos, nothing else has made it 5). I've also ran Memtest86 for about 6 hours on my memory and it had no errors, and its running well within specs right now. So, I think I've eliminated other possible problems (or have I?). I even tried tacking on +.3v to the northbridge and +.2v to the southbridge for S's and G's, but it made no difference in stability that I noticed. The system does run stable at stock (2.66ghz). I'm still trying different combinations, but so far... no luck. The system runs about 40 degrees C idle and 60 under stress (using CoreTemp and SpeedFan). All 3 of my fans are on variable speed, so they do a good job of stabilizing temperature there regardless. Is there something I'm missing? Or am I just really unlucky? I researched through a lot of other overclocks and vcores, but mine doesn't seem to fit in the same ballpark. Really, its not even the same game.

My config:
CPU: Intel C2D E6750
Motherboard: Gigabyte GA-P35-DS4 Rev 2.0
RAM: 2x2gb g.Skill DDR2-1000 PC2-8000 F2-8000CL5D-4GBPQ
PSU: 700 watt Hippro HP-W700WC3
Case: Alienware MJ-12 (aka Chenbro SR209)
HDD: RAID10 - 4x Seagate Barracuda 320g ST303204N1A1AS-RK
Case fans: Thermaltake 120mm A2018 & Antec Tri-Cool 92mm 761345-75093-6
Heatsink: Rosewill 92mm RCX-Z3
GPU: ATi Radeon x1600 512mb PCIe Pro
CD and DVD Drives: Junk

You've indicated others are getting, stable higher clocks with the E6750 (which they are), but you have yours set with a 6x mulitplier at 450MHz. It should be easier to get a higher CPU speed with a lower and more stable bus speed, coupled with a higher multiplier (6-8)

When you said, "Now I'm working through 3.2 ghz and I'm at vcore 1.4 and it still isn't stable.", I can only assume the multiplier was at 8x (400MHz), otherwise a 6x puts the bus speed at 533. It wasn't at 533, was it? Just checking...

1.51V is getting high, but if 450MHz is unstable with your CPU, no matter what the voltage, then decrease your bus speed back to 400 or 375 and go back to normal voltages (1.3500 - 1.4500 I'm guessing).

Ultimately, you may find yours is stable with 438MHz, or 442MHz, at 1.4350V or 1.4600V. They're just estimates, but you may have overshot your maximum stable bus speed and CPU voltage combination.

My point is not to work backward from an unstable set up, but work forward from a stable one.

Apart from the stock configuration, what was your last stable over clock under 3.2GHz with stock voltages?

If you've gone from stock (stable) straight to 3.6GHz/1.5000V over clock (unstable) and tried guessing from there, working your way back, that's a difficult way to do it.

In any event, leave your mulitplier on 8x. I'd start somewhere usually fairly safe, like 375MHz at stock volts and 8x multiplier. If not stable, go up to 1.4600V from stock voltage at 0.1V at a time. It's very likely you'll get some stability during this, but no guarantees.

This gives you a stable starting point beyond stock: 3.0GHz (8x) at 375MHz FSB at 1.35-1.46V.

See if 400MHz (3.2GHz) is stable at the same voltage as 375MHz. Keep going up by 10MHz until instability occurs. Then increase your voltage by 0.1 up to your maximum (1.5V or whatever you're comfortable with). By 1.4V you may be working in 0.05 increments or less.
No stability = go back to previous known stable configuration, that's about it.
Stability = go another 10MHz and see, increase voltage if necessary. So on, and so forth until no voltage increase will make it stable.

Leave the northbridge and southbridge alone until you've determined maximum bus speed and CPU volts. Too many variables if you're playing with three voltages and one frequency, plus a multiplier.

With a 8x multiplier, you'll get these bus and CPU speeds:
410MHz, 3.28GHz
420MHz, 3.36GHz
430MHz, 3.44GHz
440MHz, 3.53GHz
450MHz, 3.60GHz

Easier to get a higher CPU speed with a larger multipler and lower bus speed. Mover from a known stable configuration to a new one by changing only one variable at a time. Push bus speed until instability occurs, increase voltage until you reach your maximum or stability occurs, or go back to last stable bus speed if not able to overcome.

Ah, I was very unclear I only lowered my multiplier to x6 to test whether or not my bridge voltages and the bus speed were okay and that I wasn't hitting a wall. I've been running with x8 otherwise. I was stable at just above 3.0 ghz but after that the volts just took off and I couldn't stabilize. I didn't run a long test at 3.0 ghz, only Orthos for about 15 minutes. If I tried an overnight test, I may very well come out with some errors. Right now I'm at 3.12 testing and my vcore is already at 1.46 and it still isn't stable! It just makes me mad when so many others got 3.2 ghz with stock cooling and stock voltage and I have a decent cooling setup and can't get 3.2 at max voltage, much less stock. Granted, this is my first OC, so maybe there is something I'm missing?

As far as missing something, no, I think you've got a good grasp of what's going on.

The only other piece of information relating to instability is heat, so what is your CPU and northbridge (or mb) temp at 3.12GHz / 1.46V. Is it the 40C idle, 60C load you mention originally?

Heat is usually a good indicator of instability, or the potential for it. Increase the bus speed, but keep stock voltages, and heat is increased, if only fractionally.

For a given frequency over stock that requires greater than stock voltages, heat will increase at a disproportionately higher percentage rate than the voltage percentage rate increase. For your particular CPU at or over 3.12GHz, it may be that there is:
a) no voltage that will make it stable; or
b) the required voltage that would make it stable results in too much heat (causing instability in itself) that cannot be cooled by the current heat sink and cooling fan; or
c) some other as yet unidentified problem.

When you say "I was stable at just above 3.0 ghz but after that the volts just took off and I couldn't stabilize.", what were the idle and load temps during this? (Enough time to get temps before freeze/crash?).

If your temps were still good beyond 3.12GHz, but instability still remained, I can not think of any other reason than this is the limit of this individual processor.

That's probably not what you want to hear, so hopefully someone else can read your posts and come up the solution. Good luck.

Since all of my cooling has temperature sensors and are variable speed, the temperatures barely change. I think running at stock voltages I was at like 38 C idle and 58 C load, and then running at 1.5 I was idling at 41 C and load at 62 C. Thats using CoreTemp but checking it a couple times with SpeedFan. I guess I really don't know what my bridge temperatures are, but one of my fan's sensors is connected directly to the Northbridge and it spins in proportion with the core temperatures.

I've been working on it for about 20 hours and I was beginning to suspect that my specific processor sucks. I think I'll let Orthos run at stock for 24 hours. I only ran it long enough for it to appear stable, in retrospect...

sounds like your CPU is handicapped! Can you still RMA it? If you aren't happy with your overclock I would try RMAing and see if you get a better CPU... it might be easier than tracking down this OC culprit.

I bought it from the Egg. Will they take an RMA for such a reason? It seems unlikely. Maybe I'll stress test it for a few days at stock settings. That would be cause for RMA, yes?

What's the lowest voltage you can use and still be stable at stock speeds?

Does the board vdroop heavily like the DS3L I have?

I havent tried undervolting it yet...

and the board does have a measurable vdroop, but its not too bad. The DS4 is a pretty solid board. My power supply i pretty beefy, too. Its literally a monster.

well ive been having a weird prob near that running my 6750 @3.4 fsb @425 vcore @1.43 memory @ 425 4-4-4-12 ran orthos for an hour no problems tcase at 56 and hottest core 72 can play games for hours no prob u see when i have my system idling and screensaver comes on system freezes or hangs During screensaver or ive seen it do it while watching video's online weirdest thing its annoying .

I ran Orthos overnight two nights in a row and it blew. So, I RMA'd it. More info to come!

When you were having problems at 3.4 and 3.6, what was your memory divider?

"I researched through a lot of other overclocks and vcores, but mine doesn't seem to fit in the same ballpark. Really, its not even the same game."

Pulp Fiction FTW!

Mostly 1:1

Maybe try prime95 FFT's for about 6~8 hours (overnight). If your getting strange lock-ups like that I can almost guarantee you will recieve an error within a few hours. And assuming you do get an error, you would more than likely have to increase Vcore to get stability back. But considering you are already hitting 72C with orthos maxed, I would recommend you kick your freq down overall, or get a aftermarket HS. Later.

I'd say increase the RAM and NB voltage.

Most motherboards default the RAM voltage to 1.8 which won't be enough to run your ram at the speeds you want it too. Pump it to 2.0 or 2.1 and see what happens (should be at 2.1 anyways)

If it still isn't stable try upping the NB by 0.1 or 0.2