Lower voltage required after OC 'burnt in'? (E4300)

[Big.Wayne]

BIOS was the same, cooling was the same and certainly thermal paste isn't an issue as to get a 100MHz increase in speed needs a very large temperature drop to achieve.

Yeah I agree that thermal paste is not such a biggie but as I'm sure most people will agree thermal paste is something that does 'improve' after a few days of 'burning-in' and a few good 'thermal cycles'. I'm not suggesting that burning-in of the TIM is responsible alone for this scenario but that along with an improved user understand of how is hardware works, a few BIOS tweaks etc is a much more plausible theory. I'm pretty open minded though so don't get me wrong, I'm into chaos-theory, U.F.O, the Lost City of Atlantis etc so I'm aware that there is a lot of unexplained **** that I don't understand but I'm pretty sure from many years of experience that the 'Burn-in' concept is a web-generated 'myth'. Your evidence is a little sketchy, I'm not sure what notes you took etc so I'm afaid I will have to 'discredit' your theory, not enough evidence sorry.

You should do some research into BH5 RAM, it's pretty much legendary for running at higher speeds after being run in at 3.6V for a while. I have experienced it myself.

Well I probably would have to do some research into BH-5 'if' I was an overclock kiddy that was quite fresh-faced to the scene however having been building/tweaking computers for 12 years I saw BH-5 coming, joined in with everyone during its hayday (got a nice set of TwinMOS PC3200 that used BH-5), the only thing 'special' about that memory was it appreciated high volts and could keep 2-2-2 timings all the way to DDR500 and beyond (just).

Hey Jokester I noticed I joined the forums 1 month after you but you have double the post count and have been made a DON, lol I'm speaking to the boss

 

[Jokester]

it appreciated high volts and could keep 2-2-2 timings all the way to DDR500 and beyond (just).

When new .

My notes are here:-

First max benching overclock:-
http://forums.overclockers.co.uk/showthread.php?p=5285828#post5285828

Burnt in overclock 6 months later:-
http://forums.overclockers.co.uk/showthread.php?p=6648520#post6648520

Motherboard was DFI Ultra-D in both cases with OCZ VX RAM. When originally bought the FX57 would not bench at any voltage beyond what is shown there.

Jokester

 

[matt100]

yes I experienced a very obvious "burn in" gain too, just after I went with my second phase unit I got one of the killer x2 3800s at the start of 06, problem is everyone was spanking my clocks.

I'd heard about burn in and decided to try running my chip with 100% load for a day at stock speed with way too much voltage in it, went back to try the clocks that had failed a couple of days before and got way past them, often with less vcore.

It was too obvious to miss, a chip that won't do 2.7ghz with 1.65v doing 3ghz with 1.55 after 24 hours of massive load.

Not saying its easy to quantify but there IS an effect, I saw it.

 

[Big.Wayne]

I saw it.

 

[matt100]

I can only assume you're once again trying to be a smartass, I saw the fact that my pc was working with higher clocks at lower voltages because I sure as hell didn't smell, hear or taste it

 

[easyrider]

I can only assume you're once again trying to be a smartass, I saw the fact that my pc was working with higher clocks at lower voltages because I sure as hell didn't smell, hear or taste it

Agreed


wayne is out of his depth


yet again

 

[Big.Wayne]

Hey matt100,

I can see your having a sense of humour failure. There is a chat going here where a few folk are saying Burn-in is a real thing and I'm saying its not. If your gonna stand up and decree that burn-in is a real phenomenon then you gotta back it up with a little more than "I saw it".

Elvis is alive because a few people have seen him, the big foot exists because a few people have seen him, same for the Loch ness monster, you can see what I'm getting at here . . .

People shouldn't get so touchy if someone doesn't accept their opinion as fact, if you present sketchy finding don't be suprised if you get a big grin

I don't think there is any more to say on the subject, but if there is to be some truth in phenomenon then at the moment the best explanation has to be the 'Burning-in' or 'melting' of the thermal material between the heatsink and the CPU, also between the processor slug and the IHS.

I'm here to learn too btw

 

[easyrider]

Hey matt100,


Elvis is alive because a few people have seen him, the big foot exists because a few people have seen him, same for the Loch ness monster, you can see what I'm getting at here . . .

LOL @ wayne

Priceless

 

[Big.Wayne]

LOL @ wayne

Priceless

I know you love me really

 

[Jokester]

but if there is to be some truth in phenomenon then at the moment the best explanation has to be the 'Burning-in' or 'melting' of the thermal material between the heatsink and the CPU, also between the processor slug and the IHS.

To get a 150MHz increase in speed on an A64 needs about 15-20C in a temperature drop. Not something you're going to get from TIM bedding in. Also there's no TIM on BH5 .

Jokester

 

[z0mbi3]

To get a 150MHz increase in speed on an A64 needs about 15-20C in a temperature drop. Not something you're going to get from TIM bedding in. Also there's no TIM on BH5 .

Jokester

Just to clarify, you don't mean every 150 mhz do you? Otherwise my 600 mhz OC is a little miracle at the temps I'm getting!

 

[Big.Wayne]

To get a 150MHz increase in speed on an A64 needs about 15-20C in a temperature drop. Not something you're going to get from TIM bedding in.

Where did that info come from?

Also there's no TIM on BH5

Yeah thats true although a lot of them had heatspreaders with thermalpads underneath.

While we are here I'm just curious to know, what is the material the manufacturers splodge on top of a CPU slug before the IHS goes on? I've never removed an IHS so not sure?

 

[Corasik]

I remember some Crucial PC2100, which would run fine (at stock volts) at 266mhz without any problem, but as soon as I tried to overclock it would fail memtest within minutes, regardless of what timings and voltages I tried.

After a few months of heavy every day use, I just randomly decided to try and make it a bit faster. Switched to 333mhz (PC2700 speeds), with standard timings and it worked perfectly. Even managed to tighten up the timings. Ran perfectly at the higher speeds until I replaced the computer a year later.

So I would certainly say that computer parts can get 'better' over time. Have to admit, I've never really found a CPU thats improved, but I've been able to replicate the case of improving memory timings on several computers which had been stubborn when new.

 

[Jokester]

Just to clarify, you don't mean every 150 mhz do you? Otherwise my 600 mhz OC is a little miracle at the temps I'm getting!

No what I'm talking about is an increase in overclocking ability through temperature alone (the whole reason for using phase/ln2 etc). Reducing the temperature allows the transistors on the die to operate faster at the same voltage.

If you improved your cooling such that you got a 20C drop in temperature then it's possible you could get a 700MHz overclock or even more depending on the chip.

Jokester

 

[Jokester]

Where did that info come from?

That's from my own personal experience with phase and seeing other people results.

While we are here I'm just curious to know, what is the material the manufacturers splodge on top of a CPU slug before the IHS goes on? I've never removed an IHS so not sure?

Older A64s use a really thick paste that's basically dry, C2Ds are actually soldered on with metal.

Jokester