Overclocking (Main Thread)

TAT stands for Intel Thermal Analysis Tool, if I had to guess, but that's my take

hehe... not that, I mean what actually kills it?

darkowl72 said:Hehe... no, that's not what I meant. I was asking what actually kills it.

He meant he wanted something that would stress it even harder than Orthos.

I guess I'll try swapping out the thermal paste and maybe some other bits, then we'll see how it goes...

Just make sure you apply it in a very thin layer.

There’s a common consensus among enthusiasts that one shouldn't go overboard when cranking up the CPU voltage.
http://www.xtremesystems.org/forums/...d.php?t=100849

@Edward Peterson70/">@@Edward Peterson70
Voltage kills a CPU way faster than heat ever will.

@Twitter user
If you’re hovering just above 60°C, you aren't breaking anything, but pushing it that hard at 2.2GHz with stock voltage is definitely asking for trouble.

Harold Lopez31 said:@Edward Peterson70/">@@Edward Peterson70
Voltage does far more damage to a processor's lifespan than heat alone ever could.

That wasn't actually my point. I was just trying to show that he was pushing an Athlon to its absolute limit at 2V.

Not a reading😁 You mentioned once that sometimes it’s better not to go overboard with the voltage, so I was thinking this post could follow that same vibe: how long is it actually okay to run a processor 24/7?

It looks like this:

Gradually bumped it up to 11x255 (2805); SPD=3.20 (816)
-The TAT crashes whenever there's a load, then I can't even get back into Windows.
Same thing happens at 9x312 (2808); SPD=2.50 (780)
Keeping latencies the same: 5-5-5-15

Dropped it down to 11x250 (2750); SPD=3.2 (800)
TAT temps were sitting at 67/65, while CoreTemp showed 69/67. After about 20 minutes of use, the PC just rebooted itself.
Orthos temps stayed around 52, but after ninety seconds, the alarm goes off:
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Torture Test ran 1 minutes 22 seconds - 1 errors, 0 warnings.
Execution halted.

What exactly gave out on me?😕
What's the move here? Should I just revert to 2.6GHz where everything seemed fine?

Are you running both EIST and C1E? What’s showing up in the stress.txt file?

Harold Lopez31 said:If you disabled EIST and C1E, what does the stress.txt file actually say?

Yep. Both are off.
I can't find the stress.txt file, and searching for it isn't turning up anything either. Just my luck...

Set it to 280*10. That way, you’ll have the Democratic Party frequencies sitting as close to 800MHz as possible. Are you messing with the voltages? Check what your voltage is pulling through CPU-Z while you've got Orthus running.

Do I need to step this up gradually again, or can I just jump straight to 280*10?
..Also, when the multiplier isn't at 11x, my dad’s old PC reads the frequency wrong, and it does the same thing when I check the properties in My Computer.. I guess it might be happening to others too?

Just crack it open right away😁, but check those voltages first if you're running at a multiplier of 11; you might find the NVIDIA is set too high

e, undervolting... I wouldn't even know what the target should be, or if what I have is too high or too low. For the most part, I've got it set to manual, though I left each individual voltage on auto.
aha... just saw another post.😁

CPU-Z shows the core voltage sitting between 1.280V and 1.296V

You’ve got the exact same CPU, motherboard, and RAM combo as me. Mine runs at 333x9 on stock voltage. Just drop your multiplier down to 10 or 9 and tweak the FSB until you hit roughly 2.8GHz.

Just a heads-up though: you aren't going to notice a damn thing! Honestly, I'm fighting the urge to buy a beefier cooler, a new case, or extra fans, because I know deep down it won't give me any actual satisfaction.

Jason Hall18 said:CPU-Z shows core voltage at 1.280-1.296v

Switch those voltages over to manual. Bump the CPU up to 1.4V and see if it stays cool; if it doesn't overheat, you can always nudge it higher.
Mine was running 3DMark at about 1.5V in the BIOS, though NVIDIA brought it down to just north of 1.4V.
Just to test the limits, I even pushed it to 1.6V and, honestly, nothing happened. No issues at all.
You can also bump the memory by about 0.3V.
Not sure which sticks you're rocking, but Kingmax can usually handle an extra 0.6V (up to 2.4V).
Just a heads-up: that kind of high-voltage stuff is strictly for benchmarking—don't go running 1.6V on the CPU or 2.4V on the RAM if you plan on leaving the PC on 24/7.

Tried running at 280x10 with an FSB of 3; the system won't even boot Windows. It just keeps restarting. CPU voltage shows a normal 1.325V, I guess, and all the other voltages seem standard too since I haven't touched them.
Now I’ve switched to 240x11 (2640) with an FSB of 3.33 (799). Orthos is running fine without any issues. I think I'll just leave it there. I mean, I'm hitting speeds nearly identical to an Intel Core 2 Duo E6700, and you mentioned I might squeeze out another 300MHz later—probably wouldn't notice much difference anyway. Just hoping I didn't fry something inside.😁I suppose I'd know if I did, assuming the PC actually died.🙂