The Official OC (OverClocking) Thread!

The problem to that view is that errors often happen in small increments, don't show up until they've accrued over time. Just a bad save here and there and everything seems fine until it isn't until you have to uninstall everything and start over.

Sam I think I'm going to go with the Q9650, I need all the horsepower that I can get, and when eight core chips hit the market I will probably have to go for that. I've been communication with the developers of X264 and they've assured me that it can handle as many cores as I can throw at it.

 

Soph, I'm the guy recommending running 10 tests from linpack instead of 5. I'm all for 100% stability. Maybe you meant someone else?

 

Soph.
i understand,,,lol i just thought i said something and dont remember what it was...lol man im looseing my mind and only 39...lol

ok fella to keep you informed on my adventures in OC land..
i ran orthos all night last night a little over 8 1/2 hours everything fine. temps only hit 59c so im slap happy about that.
but i didt notice once i stopped Orthos. my vcore shot up to 1.410v. way up from the 1.36v that it was running at all night long and half the day before hand. now i did have to set my vcore in bios to 1.44325v to get it to run stable and give me the 1.36-1.38v now i know of the vdrop. but why would it run for almost 30 hours at the lower vcore. and then once done with the stress test shoot up????

as much as i want to run everyday at 4.0ghz i dont want to fry it with 1.44v so i backed it down a touch to 3.8ghz and dropped the vcore in bios to 1.375ov which is .010v higher than max volts for the CPU. and in occt,everest,cpuz.the vcore is now only 1.29v-1.32v bounceing back and forth....

like i said this morning after i turned orthos off and temps wet back to normal. even at 4.0 i was idleing at 39c

 

i like that program as much as everyone else(except Russ) lol, but i wouldnt try and run it more than 10x's that is alot of hammering if i may quote you Russ. i do think it is a little to much.even if it passes. it does make you feel better about the OC but it is alot of heat nad stress.

the main reason i brought this up is. a guy over at another site told me it wasnt worth running the IBT just 5 times try running it 50-100 times...what the hell is this guy smokeing????? he said...hell i'll copy and past the post.

yea someone is wanting a dead cpu real fast....
now im sorry thats just someone being a idiot teling someone to run that test that many times..

 

Abuzar

I think you're right. I intended that part to be for spamual but I was replying to you and sam at the same time and got things crossed.

Rob

I ran mine at 4.0 for a while and so did Mort and I decided to go with 3.8 and so did mort so not a bad choice.

On another note I just ordered a Q9650 so mu debate with self is over. It will be a week or so before I can get it installed but I think that it should work fine with my board.

 

THe InteBurnTest is NOT CONTINUOUS stressing of the components. It gives the CPU a break before each test. I think running it about 10 times is pretty reasonable. Not 50-100, that's just a death wish.

 

On air that might be true but not with decent watering cooling. Air and water can appear sometimes to have similar results and under serious stress air can't remove enough heat, but water continues to just as effective. Compressing Blu-Ray movie is just as harde on the CPU as Intelburn. Where Intelburn really stresses things that encoding doesn't is that it also goes after memory, and it's the memory that fails most often. When you see posted test results look at the amount of memory being stressed and you will see that many passing systems used lower amounts.

 

abuzar1,

Different kind of stability than what Sophocles was talking about. It's a fact that files do get corrupted over time. I consider myself fortunate that my XP-Pro install on my last drive lasted two years! At the end there were so many glitches, yet it would pass 12 hours of Prime95. It's not always the OC! You just don't know in advance what will happen! Runs like a jewel now. It ran good up until about two weeks ago. LOL!! Believe me when I say that hardware stability is easier! LOL!!

Because of the way the problems occured with mine, I was pretty sure it was software related!

Best Regards,
Russ

 

Vcore will rise when test finishes due to loadline calibration. Essentially, LLC detects when your cpu is being loaded, impending vdroop and raises the core voltage higher. When the load stops, it cuts it back, so the process isn't instantaneous - for a moment, the test has stopped and there's no vdroop, but the higher core voltage is still being applied, thats why the voltage jumps at the end of the test (I think)

 

Perhaps I was unclear. When a piece of metal is heated the heat is conducted through its entire surface area which has a temporary cooling effect at the source. The mass along with the conductivity of the metal will determine how much heat is drawn away from the core and how fast. Eventually the entire mass of the heat sink is heated and it is less able to conduct or draw heat from the core which is why a fan is used to pass air through its radiator, to transfer heat away from the metal. Under intense stress the core will continue to transfer heat to the heat sink and if the amount of energy going in is greater than that being removed by air then the core will get hotter. Water has much greater mass than air and because it is moving through a metal block across the core continuously there is effectively even more mass coming into contact and more heat is being transferred than is being generated. The core will continue to heat to a point but then it will level and stop rising. I've seen some of the best air coolers allow for heat increases that are more than twice that of water.

 

that makes sence, but the volts before stress and during stress were between 1.36-1.39v never any higher. but once i stopped orthos it went up to 1.41+.... i dont know... it is all fine now and running at its retirement speed of 3.8ghz...lol

 

It should only go to 1.41 briefly, then cut back to 1.38 once LLC realises the higher voltage is no longer required.

 

i always reinstall my OS every yaer or so, i guess just out of habit, and my chip only needs to last a year, maybe less if i switch to i7 early enough.

i know im being stubborn, but im sure you lto guess by now, when it comes to components, i usually am

i like to prove people wrong LOL

 

it might do that Sam. i know i watched it for about 15 min's and it never droped back down... but there aint no worries as i dropped it a bit to 3.8ghz and volts are great.

i need to find out if i can adjust the FSB volts with this board.i has all the otehr options everyone has talked about except the FSB volts.

the speed of this thing compared to the other faulty CPU is unreal. i cant get over how fast programs and such open up includeing windows..

 

Rob

I find that around 1.36 gives a pretty stable 3.8 GHz. With vdroop you'll get this.

 

i have my vcre set now @1.3725v i think it is. and with the vdrop its inbetween 1.28-1.32v running at 3.8ghz..i might try and set it at 1.35v and see if i can keep it stable

 

I've found that my system will run most things at a lower vcore but when it comes to stress testing things tend to go wrong. I also have my memory set at 2.10 so my settings are about where I like them across the board.

 

Soph you think i should try and drop my mem volts? im at 2.2v now there rated voltage..

on my vcore. i tried to set it at1.33?? and it wouldnt get stable. so im thinking between 1.34 and 1.35v i might be able to get it stable.

 

Rob you might look at these voltages and see what happens. I looked thru all your pix again and didn't see the "FSB" indicator either.

ICH = SB 1.5v => 1.55v-1.6v
MCH = NB 1.1v => 1.15v-1.25v
cpu. PLL 1.5v => 1.55v-1.6v

hth in some way....

...gm

If someone can enlighten us both that would be nice. tia....

here's his BIOS pix

 

I just read like the last three pages of this thread, and I have to say : YOU GUYS ARE CRAZY!!!!! (in a good way )

I mean the amount of technological/computer know-how you guys have is amazing.

Maybe I'll be the same someday ...........

BUT to gain knowledge you need access to it which is kind of difficult on dial-up P), but I'm working on it