Learned a New way to pick out a good chip!

[aigomorla]

Okey guys i wanted to share this with you, because he sounds like he's right.

On my quest to look for that golden QX chip i ran into this thread where hipro made this comment:

Originally Posted by hipro5
Thanks guys.....

This number on the PCB of the CPU is indicating (from my obsevations all these years that I'm "focusing" on it) the "place of the wafer" it was cuted....Propably the LOWEST (closed to 00) the TWO LAST numbers are, the "centered of the wafer" it was cuted...
For example: LAST TWO numbers.....000, 100, 200, 300, 400, 500, 600, 700 and so on....
FIRST 15 numbers are the best....for example2: 012, 005, 009, 104, 113, 202, 215, 205, 314, 304, 406, 412, 601, 609, 712, 703, etc....
These 2 LAST numbers should be closest to "00" LIKE: 001, 713, 404, NOT 478, 656 and so on....

At least that's what I have observed by searching all these years....


Now you guys are going to ask what this PCB number is. To most of us it wont mean much, but to the people looking for that top 2% it plays a hugh role:

The number Hipro is talking about is this:
http://i125.photobucket.com/al...aigomorla/IMG_0884.jpg

see where it says at the bottom 4A0004. According to hipro via PM's I asked him for a better explaination and he responded with this:

Though L740 are "bad" clockers, the digit of yours 0004 seem to be a VERY good piece of sillicone and I think that numbers like 0001 to 0010 wasn't given "out" by Intel that easy.....


So the last 2 numbers you want to get as close to 00 as possible. And at first i thought he was crazy about this prediction til i saw this:
http://i125.photobucket.com/al.../p73/aigomorla/OMG.jpg

Its not prime stable yet, but its the lowest voltage required for windows stable operation. A few more voltage bumps and it should get prime stable. I'll let you guys know more when i finish rebuilding persephone and i get to fully unlock it.

I hope this info helps you all in finding that golden bin chip! Remember we now have 2 set of numbers to look at and not just 1.

 

[JAG87]

It's BS.

An intel employee on xs said that the number in question is a randomly generated PCB serial number that has absolutely nothing to do with where the wafer was cut.

And since the batch number is still not a reliable method for finding a good clocker, its still all luck of the draw.

 

[aigomorla]

yes JAG if you think its BS please show me.

As i said i was skeptical at first when he said this, but my evidence to his claims are pretty high.

 

[JAG87]

http://www.xtremesystems.org/f...p=2766054&postcount=53

the thread below also has a ton of controversy about those numbers. Kingpin himself asks Hipro5 why is he telling us this bs. Its hard to tell if its sarcasm or not, but there is lots of other proof that supposed bad chips clock well, and supposed good chips clock bad.

http://www.xtremesystems.org/f...howthread.php?t=167336

and to conclude, my chip's number is 504 and its nothing special

 

[Idontcare]

Originally posted by: aigomorla
As i said i was skeptical at first when he said this, but my evidence to his claims are pretty high.

Anecdotal evidence need not be correct 100% of the time in order for it to be useful some of the time.

I for one appreciate the info Aigo.

If it's 100% BS then the worse that happens it you have no better or worse chances of getting a great chip than if you hadn't bothered to try in the first place.

If it is viable information and the odds are truly higher than 50/50 that these numbers are where the good chips can be found then I you at least increase your odds.

So there's no harm done if its wrong, but it could be an edge for those of us wishing to increase our chances of having an edge.

Originally posted by: JAG87
It's BS.

Lolz, JAG bags himself another threadcrap.

If you don't like what Aigo puts in his own thread then move along.

This forum isn't called "the world according to JAG".

 

[aigomorla]

actually i take JAG's comments quite honestly.

he's very straight forward with me, lacks emotion when saying im wrong, he just gets to the point.

That link is indeed crediable, however, maybe if we compare more processors.

I'll pull my G0 off, as i thought it was a decient overclocker, and read out the patch. Maybe if we compare enough samples we can come to our own conclusion.

And i'll take anything that will increase my odds even 15% on finding a better chip.

 

[Markfw]

Well, I just got something better to do with my spare time then pull off 12 heatsinks to check the numbers.....(I just got a squeeze) So you will have to do this one without me.

 

[JAG87]

Originally posted by: Idontcare

Originally posted by: aigomorla
As i said i was skeptical at first when he said this, but my evidence to his claims are pretty high.

Anecdotal evidence need not be correct 100% of the time in order for it to be useful some of the time.

I for one appreciate the info Aigo.

If it's 100% BS then the worse that happens it you have no better or worse chances of getting a great chip than if you hadn't bothered to try in the first place.

If it is viable information and the odds are truly higher than 50/50 that these numbers are where the good chips can be found then I you at least increase your odds.

So there's no harm done if its wrong, but it could be an edge for those of us wishing to increase our chances of having an edge.


Absolutely.

Originally posted by: JAG87
It's BS.

Lolz, JAG bags himself another threadcrap.

If you don't like what Aigo puts in his own thread then move along.

This forum isn't called "the world according to JAG".


I dont think you know what a forum is. A forum is a bunch of people sharing their opinions on a subject, whether they are positive or negative, just like you did above.
There is nothing wrong with me stating my opinion that this is BS, as long as I do it in a civilized way.

 

[Markfw]

There is nothing wrong with me stating my opinion that this is BS, as long as I do it in a civilized way.

Exactly !

 

[bunnyfubbles]

Unless you had several unbiased samples of low numbers achieving great clocks/low voltage and high numbers struggling with high clocks and requiring a lot of voltage, I'm going to be skeptical as well. It'll be something to keep in mind, and since its out there we can now keep an eye out for it and do our own research into the matter.

 

[PCTC2]

I'll have to pull my E6400 out...

 

[aigomorla]

Originally posted by: bunnyfubbles
Unless you had several unbiased samples of low numbers achieving great clocks/low voltage and high numbers struggling with high clocks and requiring a lot of voltage, I'm going to be skeptical as well. It'll be something to keep in mind, and since its out there we can now keep an eye out for it and do our own research into the matter.

hey i was skeptical at first as well.

im mearly sharing a finding that i have with you guys.

I am not saying this is fact, cuz who knows it might not, but i am letting you guys know what i saw from my obseravtion.

 

[sonoran]

Well, I've been at Intel over 7 years, and I have never seen anything indicating that a chip being from any specific location on a wafer makes it more likely to be a high-clocking chip. For that matter, I've never seen anything indicating that there are more defects farther from the center. And the way a silicon wafer grows (it's one BIG crystal, growing at the molecular level), I can't imagine why any portion of the wafer would be more or less likely to have defects than another. I know this rumor's going around the boards, but I have yet to see one single *real* study showing it's true. If someone has real technical articles to this effect, I'd be curious to see them.

If anything, I'd expect chips that are farther from the alignment edges of the wafer to have more defects. Fact is the wafers have to be loaded and imaged by the litho equipment multiple times, and the precision of alignment needed is incredible. I could readily imagine that chips further from the alignment edges would be more prone to defects if there were the slightest variance in the litho equipment, or room temp, or vibration, or...you get the idea. But here again, I've never seen any studies on this.

So if anyone has links to real technical info, I'd love to see it. For now, this sounds like so much snake oil to me. But as another poster said, looking for chips with these numbers isn't likely to hurt you any - so knock yourself out if it makes you happy.

* Not speaking for Intel Corp *

 

[dmens]

The center dies are simply more representative of the average behavior of a lot, not the fastest. Golden parts come from specific wafers, not the center of every wafer.

I have no idea what those numbers are, but I really doubt there is some kind of code that indicates more margin than other parts. Looking for a "golden part" without information from a tester is pretty much a total waste of time.

 

[bryanW1995]

Originally posted by: aigomorla
actually i take JAG's comments quite honestly.

he's very straight forward with me, lacks emotion when saying im wrong, he just gets to the point.

That link is indeed crediable, however, maybe if we compare more processors.

I'll pull my G0 off, as i thought it was a decient overclocker, and read out the patch. Maybe if we compare enough samples we can come to our own conclusion.

And i'll take anything that will increase my odds even 15% on finding a better chip.

I'll check the numbers on my X3350 when it arrives and my e6750 when I remove it.

@dmens: so if 3 or 4 people have chips from the SAME WAFER and they are all great oc'ers then that could be an indication that the entire wafer was superior, right?