Thermal paste

[erocker]

I say let Dead Space 2 be the judge of that.

[anas]

yes hop it wont be dead cpu 2 hehe you want a screeny with fraps when i install it tomorrow ?

[Velvet Wafer]

Originally Posted by anas

yes hop it wont be dead cpu 2 hehe you want a screeny with fraps when i install it tomorrow ?

its not the matter of death, but instability erocker played on i guess.

Originally Posted by Bo$$

read, about 3-4 people including me mentioned a linpack tester...

I did a search for IBT specifically - and as I've already said if you read my post, this is not your ordinary linpack tester.

[anas]

Originally Posted by Velvet Wafer

its not the matter of death, but instability erocker played on i guess.

time will tell , i will not be shy to speak about my failure but this stock fan aint bad at all .

one thing i found tho i wounder if its true or its just my ignorance :

someone told me that raising fsb is like raising
multiplier = same cpu freq

but with my tests i found that raising fsb needs more vcore than raising
in multiplier

ex : i get to 4ghz with 421x9.5 at vcore 1.216 and a bit of ram increase

but i need 1.3 vcore for 450x9 + more ram speed = crashes , more heat and power use, and also degrading hardware . can i be so wrong ?

[Velvet Wafer]

Originally Posted by anas

time will tell , i will not be shy to speak about my failure but this stock fan aint bad at all .

one thing i found tho i wounder if its true or its just my ignorance :

someone told me that raising fsb is like raising
multiplier = same cpu freq

but with my tests i found that raising fsb needs more vcore than raising
in multiplier

ex : i get to 4ghz with 421x9.5 at vcore 1.216 and a bit of ram increase

but i need 1.3 vcore for 450x9 + more ram speed = crashes , more heat and power use, and also degrading hardware . can i be so wrong ?

you could start by lowering the memory divider, so a higher FSB will produce not as much memory clock. adding voltage should be the last thing to do, when you overclock

[fullinfusion]

Like Velvet said... +1 on that.

You taking one thing and turning it into 2 things and trying to control both with problems.

Take one step at a time my friend

[de.das.dude]

oh and as to answer your first queston if your thermal paste will be damaged, and also to test a chipset cooler 's efficiency,
i took my 35W soldering iron. put a dab of TIM on the chipset cooler and touched the soldering irons tip to it. soldering iron was completely heated up beforehand.

so?? nothing happened, the cooler worked perfectly, dissipating the 35W and keeping the itself cool.
as for the TIM, it kinda got dry after a couple of mins.

also, there is a temperature range already provided by the manufacturer on the TIMs package, stick to it (its impossible not to :P )

oh i should mention i tried without the fan on the cooler to check taht thermal paste was actually working... and it went to 60C in 10 secs without the fan. temp sensor?? nerves on my finger tips :P

[anas]

450 fsb means 900 mem divder as minimum .plus you didnt get the main quiestin ,its about
fsb and multiplier :

someone told me that raising fsb is like raising
multiplier = same cpu freq

but i found that lower fsb +higher multiplier = less voltage needed and more heat
and the other way around . is that true or its my imagination ?

[de.das.dude]

Originally Posted by anas

450 fsb means 900 mem divder as minimum .plus you didnt get the main quiestin ,its about
fsb and multiplier :

someone told me that raising fsb is like raising
multiplier = same cpu freq

but i found that lower fsb +higher multiplier = less voltage needed and more heat
and the other way around . is that true or its my imagination ?

well.... FSB kinda craps your mem settings, so you need to set memory timings and speed everytime you OC with FSB. i use fsb to OC cuz i dont have and upward unlocked multiplier.

[newtekie1]

Originally Posted by twilyth

I did a search for IBT specifically - and as I've already said if you read my post, this is not your ordinary linpack tester.

Nah:



Linpack is Linpack, IntelBurnTest on Max doesn't heat up the CPU any more than OCCT Linpack on Max. I'd agree with you if you were talking about the normal OCCT stress test, but not talking about OCCT Linpack. It doesn't matter if it is IBT, or Linx, or OCCT they all heat up the CPU the same because they are all doing the same calculations.

Originally Posted by newtekie1

Nah:
http://img.photobucket.com/albums/v2...tekie1/IBT.png
http://img.photobucket.com/albums/v2...ie1/occt-1.png

Linpack is Linpack, IntelBurnTest on Max doesn't heat up the CPU any more than OCCT Linpack on Max. I'd agree with you if you were talking about the normal OCCT stress test, but not talking about OCCT Linpack. It doesn't matter if it is IBT, or Linx, or OCCT they all heat up the CPU the same because they are all doing the same calculations.

You're simply wrong. I've seen it on several machines. If you think I'm lying fine. I don't really give a shit what you think.

edit - sorry being defensive but I stopped posting here (outside of WCG) partly because I got tired of the snarky-ass comments I had to put up with. Having said that, I'm not inclined to prove anything to anybody. If you don't want to take my word for something then you don't know me very well. However that fact alone does not incline me to waste the time necessary to prove my point. If I have absolutely nothing better to do tomorrow, I will think about running tests with OCCT and IBT.

I would also point out that your scn prt show you didn't even complete one IBT run, so it's not like your prts amount to anything even resembling "evidence".

[Laurijan]

Originally Posted by twilyth

You're simply wrong. I've seen it on several machines. If you think I'm lying fine. I don't really give a shit what you think.

edit - sorry being defensive but I stopped posting here (outside of WCG) partly because I got tired of the snarky-ass comments I had to put up with. Having said that, I'm not inclined to prove anything to anybody. If you don't want to take my word for something then you don't know me very well. However that fact alone does not incline me to waste the time necessary to prove my point. If I have absolutely nothing better to do tomorrow, I will think about running tests with OCCT and IBT.

I would also point out that your scn prt show you didn't even complete one IBT run, so it's not like your prts amount to anything even resembling "evidence".

How´s about meeting outside of the Saloon at 12am?

Originally Posted by Laurijan

How´s about meeting outside of the Saloon at 12am?

Hey dude. How's it going. Check your god damned pm's and vistor messages once in a while.

Yeah, I must be a little wound up. 4:23am EST and I'm still wired. You'd think the tranks, anti-psychotics and the rest of my nightly cocktail would have SOME affect. But my neurons just keep partying at my mental redline.

[anas]

Originally Posted by Velvet Wafer

you could start by lowering the memory divider, so a higher FSB will produce not as much memory clock. adding voltage should be the last thing to do, when you overclock

lowest mem divider for 450 fsb is 900mhz and my poor kingston kvrs dosnt like that .
at 421 fsb they go like 840 and 1.9v dram so they are more stable

[anas]

a stress test , why you didnt say that from the start .here you go :



i will not hammer my cpu with the other stuff , live and let die

[newtekie1]

Originally Posted by twilyth

You're simply wrong. I've seen it on several machines. If you think I'm lying fine. I don't really give a shit what you think.

edit - sorry being defensive but I stopped posting here (outside of WCG) partly because I got tired of the snarky-ass comments I had to put up with. Having said that, I'm not inclined to prove anything to anybody. If you don't want to take my word for something then you don't know me very well. However that fact alone does not incline me to waste the time necessary to prove my point. If I have absolutely nothing better to do tomorrow, I will think about running tests with OCCT and IBT.

I would also point out that your scn prt show you didn't even complete one IBT run, so it's not like your prts amount to anything even resembling "evidence".

Yes, and yet IBT had been running longer than OCCT had been. When set on maximum it takes forever to complete a run of IBT, you should know that if you're so experienced with it, and being 3AM and needing to be up at 7AM I didn't much feel like sitting there for another hour+ letting both run. I've had plenty of experience with IBT, and OCCT heats up the CPU just as much. Sorry if I don't take your word for it, but my evidence and experience says otherwise. I'll gladdy take some evidence from you though showing differently. And the reason I prefer OCCT is that it has a built in thermal stop, so I can set it to run and walk away without worrying about it, unlike IBT.

And there was nothing snarky-ass about my response. If you take every post that disagrees with you as snarky-ass, then I'll be glad to see you leave again.

Originally Posted by newtekie1

Yes, and yet IBT had been running longer than OCCT had been. When set on maximum it takes forever to complete a run of IBT, you should know that if you're so experienced with it, and being 3AM and needing to be up at 7AM I didn't much feel like sitting there for another hour+ letting both run. I've had plenty of experience with IBT, and OCCT heats up the CPU just as much. Sorry if I don't take your word for it, but my evidence and experience says otherwise. I'll gladdy take some evidence from you though showing differently. And the reason I prefer OCCT is that it has a built in thermal stop, so I can set it to run and walk away without worrying about it, unlike IBT.

And there was nothing snarky-ass about my response. If you take every post that disagrees with you as snarky-ass, then I'll be glad to see you leave again.

Yes, IBT does take quite a long time - that's sort of the point. If you watch OCCT, it cycles up and down during each run so you never really stress the chip in the same way. IBT goes full throttle and stays there. THAT is why you get much higher temps with IBT than OCCT. Only towards the very end of each run with IBT do you see the temps drop - and even then, not by a huge amount. As soon as you see that, within a minute or two you're off on the next run, again at full throttle.

I'll make you a deal. Run IBT on max and free up as much memory as possible before hand. Watch the temps continuously for the second or third run and get a feel for when you hit the max. Then, on the next run, when you hit the high temp, THEN do your screen shot.

Do at least 5 runs, but preferably 10 - and yes, it does take a long time.

Then run OCCT and use any screen shot you like. If the highest IBT temp is not at least 5C hotter than any OCCT temp, I'll waste the time to do the same test myself on one of my 1090T's and prove that you're wrong.

I think that's a fair deal. We can even make a wager if you like.

[MxPhenom 216]

Originally Posted by n-ster

IMO AS5 sucks... sucks big sweaty balls

MX-2 or MX-3 or OCZ Freeze

MX-4 is better then all three of those. I hated MX-3 it didnt not spread well. OCZ and MX-2 spread easily cuz they are thinner but i like MX-4 a lot its just the right consistency

Also stress test your CPU with Prime95 and or 50 rounds a IBT/LinX like mentioned above

[micropage7]

yap, just remember thermal paste has warm up time so after it applied it may not show its best performance, after warm up time the thermal paste would show you the real performance so to take objective verdict do the test after warm up time

[newtekie1]

Originally Posted by twilyth

Yes, IBT does take quite a long time - that's sort of the point. If you watch OCCT, it cycles up and down during each run so you never really stress the chip in the same way. IBT goes full throttle and stays there. THAT is why you get much higher temps with IBT than OCCT. Only towards the very end of each run with IBT do you see the temps drop - and even then, not by a huge amount. As soon as you see that, within a minute or two you're off on the next run, again at full throttle.

I'll make you a deal. Run IBT on max and free up as much memory as possible before hand. Watch the temps continuously for the second or third run and get a feel for when you hit the max. Then, on the next run, when you hit the high temp, THEN do your screen shot.

Do at least 5 runs, but preferably 10 - and yes, it does take a long time.

Then run OCCT and use any screen shot you like. If the highest IBT temp is not at least 5C hotter than any OCCT temp, I'll waste the time to do the same test myself on one of my 1090T's and prove that you're wrong.

I think that's a fair deal. We can even make a wager if you like.

IBT cycles the same way. You've never noticed that? When it finishes a run, it cycles down and back up again, just like OCCT Linkpack(because their the same test...).

No need to take a screen shot at the exact moment of the maximum temp, luckily Realtemp records that already and shows it to you.

IBT:


OCCT:


OCCT gives the same temps after 7 minutes of testing, and I let IBT run for 30 minutes after 7 runs gives the same temps. I'm waiting for your evidence. Start wasting your time, and I do mean wasting you time, because I doubt you will be able to show any different.

Originally Posted by micropage7

yap, just remember thermal paste has warm up time so after it applied it may not show its best performance, after warm up time the thermal paste would show you the real performance so to take objective verdict do the test after warm up time

AS5 is really the only paste that still has a cure time, all the other pastes don't require any cure time.

Originally Posted by newtekie1

IBT cycles the same way. You've never noticed that? When it finishes a run, it cycles down and back up again, just like OCCT Linkpack(because their the same test...).

No need to take a screen shot at the exact moment of the maximum temp, luckily Realtemp records that already and shows it to you.

IBT:
http://img.photobucket.com/albums/v2...ekie1/IBT4.png

OCCT:
http://img.photobucket.com/albums/v2...kie1/OCCT4.png

OCCT gives the same temps after 7 minutes of testing, and I let IBT run for 30 minutes after 7 runs gives the same temps. I'm waiting for your evidence. Start wasting your time, and I do mean wasting you time, because I doubt you will be able to show any different.



AS5 is really the only paste that still has a cure time, all the other pastes don't require any cure time.

As I said - 5C difference. **edit** well, 4 and change I guess.

It's not my fault you don't know how to use these utilities.

[n-ster]

Originally Posted by twilyth

You're simply wrong. I've seen it on several machines. If you think I'm lying fine. I don't really give a shit what you think.

edit - sorry being defensive but I stopped posting here (outside of WCG) partly because I got tired of the snarky-ass comments I had to put up with. Having said that, I'm not inclined to prove anything to anybody. If you don't want to take my word for something then you don't know me very well. However that fact alone does not incline me to waste the time necessary to prove my point. If I have absolutely nothing better to do tomorrow, I will think about running tests with OCCT and IBT.

I would also point out that your scn prt show you didn't even complete one IBT run, so it's not like your prts amount to anything even resembling "evidence".

blame tpu members who like proof and evidence over believing some random guy on internet. ofc you shouldn't justify yourself all the time as ppl should do their own research by trying it themselves or even google, but in this case, proof can be asked without it being snarky assed

however I myself have always wonder which is true. but to me, I've never heard of OCCT beating IBT, only matching it, so I just chose IBT logically

Originally Posted by n-ster

blame tpu members who like proof and evidence over believing some random guy on internet. ofc you shouldn't justify yourself all the time as ppl should do their own research by trying it themselves or even google, but in this case, proof can be asked without it being snarky assed

however I myself have always wonder which is true. but to me, I've never heard of OCCT beating IBT, only matching it, so I just chose IBT logically

To show how out of touch I tend to be, I thought it was common knowledge that IBT was the harshest stability tester. I only started using it about a year ago after years of using prime95. I was shocked at the temps it produced. After a while, I just assumed that anyone who tried it would see how much it could stress a system and automatically pick it over whatever else they were using. So when I got to this thread and didn't even see it mentioned, I couldn't belief it and had to double check.

[newtekie1]

Originally Posted by twilyth

As I said - 5C difference. **edit** well, 4 and change I guess.

It's not my fault you don't know how to use these utilities.

http://i181.photobucket.com/albums/x...h/cab13f2a.png

http://i181.photobucket.com/albums/x...h/c4b74db9.png

And you want to say I'm the one making snarky-ass comments? Sorry for asking for some evidence, I'm sure you'll think this was a snarky-ass comment as well, because I'm sure every comment that doesn't agree with you entirely is considered snarky-ass, right? Let me break it to you, you'll find a lot of snarky-ass comments if that is your standard, so you might as well leave again, you won't be missed.

Yeah, lets see some Realtemp numbers, not speedfan that doesn't keep track of what the actual max was. Because the temps can jump 3-5° in a matter of a second with both tests, which is why I use realtemp to tell me what the actual max reached was.