Sinner

It's always an interesting discussion; someone's beliefs (based on tests, readings, manufacturers' specifications etc.) versus someone else's experience.

I'll trust the experience, most of the time. And it tells me that this

is just not true.

It might happen, but WILL not.

__________________
I read, much of the night, and go south in the winter.

DuckieHo

Can you prove that it didn't degrade? Can you prove the electromigration didn't accelerate drastically and shorten the expected lifespan of your CPU?

__________________
To answer most of your questions: (1) a fridge cannot cool a PC (2) 64-bit OS for over 3.4GB (3) If a PCIe card fits, it should work (4) Resolution, not screen size (5) If you have a question, it is not news (6) Report, not respond to Spam (7) Single-Rail/Non-Modular PSUs are not always better than Multi-Rail/Modular

gurusan

Well, it still clocks better than ever...and I'm willing to bet that in months or even years it will clock just the same.

__________________

RyanRacer48

gurusan that very well may be the case.

However there is NO answer to the question. You may have lucked out and own a chip that can run 1.5 all day long 24/7 and never hurt the cpu. You also may have horrible luck and realize 6 months to a year down the road that your chip is running sluggish, previous stable o/cs are no longer viable etc.

Push what you can out of it, just keep a fund going to cover your replacement chip should it fry. but dont take that as a rude comment as it applies to each and every overclocker on this forum NOT excluding myself.

Say your buying a Intel CPU w/ a stock running speed of 3 ghz. You have purchased a chip Intel has tested and verify it will run cool and stable that that speed REGARDLESS of VID. When you want to push it beyond those stock limits you are taking a chance either way you look at it.

The further you push a chip the higher the chance gets of it failing. Whether that failure is simply a dead chip or a degraded chip it is still failure.

You overclock at your own risk, if your overclocking ESPECIALLY voltages its key you understand these things.

Personally if someone wants to run their C2D at 1.6v . and they are stable w/ reasonable temps. Half at it! You may just end up getting a forced upgrade in a year!

__________________
" A lot of things wouldn't get done if it weren't for the last minute. "

http://www.overclock.net/sale/588228-worldwide-new-hdmi-xbox-360-iz3d.html

MY BUILD THREAD - BONEYARD

10acjed

Big rep to Pizza for the research and effort here.

These data sheets are very useful to understand the "Absolute min/max Ratings" and are by no means "the law", if you feel the desire to run 1.45 - 1.5v to your 45nm cpu, by all means do so and please keep track of how long you have been doing so. It is how things are figured out.

If you are like myself and alot of others here, we spend most of our xtra $ on our PC already and OC to squeeze the most performance out of it while still running safely. So we dont have the means to replace cpu's for the good of overclockers everywhere.


Another thing everyone must take into account is that intel says that "normal operational limits" of a Q9550 is @ 2.83GHz, anything more than that is "outside" of the parameters and will cause premature failure. They do not increase the FSB to achieve higher clock freq, they increase the bus multi.

Remember these data sheets do not say anywhere that it is safe for us to run our cpu's at 100Mhz past its default freq. Yet here we all are pushing as far as we can...

__________________

ericeod

I've noted 3 CPUs that have been killed from high vcore and LLC, but members seem to dismiss it as user error and take it with a grain of salt.

These white papers have been around since the 45nm release, but the i7 has just been added.

See post #68:
Oc/benchie: E8400 - Page 5 !!!

See the after affects of running 1.45v vcore on the same chip as above:
E8400 Crapped out! YAY!!!!

Note that the QX9770 runs at a 400 FSB and is the same core as a Q9550, so Intel does run the cores at 400 FSB.

QX9770 Intel Specs:
http://processorfinder.intel.com/det...px?sSpec=SLAWM

__________________
eVGA 680i Overclocking guide .......... ASUS P5E/Maximus Thread.......... 4.4GHz CPU-Z

10acjed

I never realized that about the QX9770 running at 1600 as apposed to the 1333 of its siblings. Interesting...

As for the Dead E8400 running at 1.417v, that is above the max voltage after Vcc overshoot of .050 would equal it being 1.467v every time it goes from partial load to idle.

Another thing everyone is not looking at is the FSB VTT voltage limits being 1.45 as well, and withing the same dangers of .050 over shoot. So running over 1.4v on the FSB could cause damage as well..

Personally I stay a bit below the max, my Q6600 runs at 1.46v and if I was on a Q9xx I would not run it past 1.38 for 24/7 use on either the FSB or cpu core.


I am just going by what intel says, and they say in black and whit that the max voltage is 1.45v. AGAIN over shoot is .050v so that would drop it to 1.40v, similar to the 65nm being 1.55v but we stay below 1.50 for safe 24/7 use.

__________________

PizzaMan

Here lately I've been questioning everything I've been reading about voltage and temperature limitation. Mostly, it's the release of i7, its specs from Intel vs the results from RL users. Intel states a 1.55v and 68C max, but most are limiting them selfs to 1.37v and are running 80 and 90C with 1.37v. That makes 1.55v seem just plain crazy. Either we have been looking to hard at the data Intel gives up for a temperature monitoring point(Tcase) that they don't give a sensor for. Instead we get sensors for Tjunction which are not temperatures at all. They are actually a point of measurement away from TJmax. Though, software creators of programs like CoreTemp assume a TJmax and make a guess at your temperatures. But Intel doesn't mention TJmax anywhere in their data sheets. I'm sure Intel is doing some of these intentionally.


Anyways as regard to temps, monitoring your distance from TJmax is really only the most accurate way to do it. Then you most ask yourself, how far away from TJmax do I wont to stay?

__________________

____________________________
.:|Official|Volt|Mod|Squad|Member|:.

GTL testing and tweaking for all 7 series nVidia boards

Blameless

Any increase in voltage, clocks, or temps equals greater electromigration and shorter CPU life.

However, how much shorter, and wether the difference matters, is a much more difficult question.

The temps reported by realtemp, coretemp, everest, or anything else that reads from Intel DTS have exactly jack of all to do with the thermal limit Intel states. Intel's thermal spec is based on the assumption that TCASE is being measured, not tjunction.

You cannot read TCASE unless you pysically mill a trench in your IHS and insert a thermocouple.

Also, 1.375v is the VID range. Theoretically, if you do nothing do limit vdroop, and you set 1.375v in the BIOS, you are within the voltages that Intel themselves consider safe.

1.55v is the absolute max rating, which takes into account momentary surges and spikes. No one at Intel, and no OCer not on sub-ambient cooling would consider this a safe 24/7 voltage. Yes, Intel is usually pretty conservative with their voltage ratings. Still running at 1.55v is, in all probability, going to dramatically reduce the life of the chip (it may still last quite a while though, depending on how much its run at it's limits).

Depends on your voltage and clocks, and how long you want the chip to last.

Intel's warranty is 3-years. It's a pretty safe assumption that the vast majority of their chips will last at least that long if run at stock speeds, at the top of the VID range (again, before vdroop/voffset), right below the TJmax.

To keep the rate of electromigration the same while OCing, or over volting beyond this, you'll need to keep the CPU colder.

Triscuit

thanks! i was looking for something like this