When will the core2 chips be cracked?

[HeXploiT]

I cannot believe core2 chips have been on the market for as long as they have and no one has figured out how to unlock the multiplier. Surely everyone knows that if the core2 can be cracked it'll make for a hayday for us overclockers. So are we any closer to cracking this chip than we were a year ago?
Looking for any light anyone has to shed on the subject.

 

[Lord Banshee]

What other CPUs have been cracked in terms of unlocking higher multipliers?

 

[HeXploiT]

Pentiums, Pentium II's, Pentium III's, athlon xp's...don't know about pentium IV's as I never owned one.
Most cpu's throughout history.

 

[Lord Banshee]

Well i never heard such a thing for of P4, a64, core, PM, or core2duo.

 

[Markbnj]

Maybe nobody sees the point anymore .

 

[Lord Banshee]

Or intel and amd just laser cut them now as if they didn't why sell EE editions to the only people that would buy them would be the same people that would mod the CPUs lol

 

[HeXploiT]

Originally posted by: Markbnj
Maybe nobody sees the point anymore .

Obviously there's a point or they wouldn't be locked.
I'm give my little finger to spare my northbridge some heat.

 

[NanoStuff]

Originally posted by: Lord Banshee
Or intel and amd just laser cut them now as if they didn't why sell EE editions to the only people that would buy them would be the same people that would mod the CPUs lol

Laser cut to prevent overclocking? How does that work?

They should just laser cut the whole thing so people would have to buy more.

 

[Lord Banshee]

Well, maybe not laser cutting but maybe some special registers that get burned in at testing. My guess the only way to revert such changes is to have the some of the equipment that does that kind of stuff.

But i do not think the only reason to do is not for overclockers, but so OEM dealers don't sell cheaper processors that work at higher multipliers.

 

[BrownTown]

I'm like 95% sure that they have fuses on the multiplier which are burnt in during fabrication for non-EE chips, it is physically impossible to unlock these chips.

 

[firewolfsm]

Originally posted by: BrownTown
I'm like 95% sure that they have fuses on the multiplier which are burnt in during fabrication for non-EE chips, it is physically impossible to unlock these chips.

Unless you can guess which of the hundreds of millions of transistors locked your chip

 

[BrownTown]

Originally posted by: firewolfsm

Originally posted by: BrownTown
I'm like 95% sure that they have fuses on the multiplier which are burnt in during fabrication for non-EE chips, it is physically impossible to unlock these chips.

Unless you can guess which of the hundreds of millions of transistors locked your chip

IT don't even matter if you know exactly where the problem is at, there is no possible way you could fix it. Just in theory you would have to start by debonding the chip from the packaging then etching off all the inter-level dielectric then etch away or deposit the correct vias to undo the fuses then regrow all the metal layers above etc..

IE: impossible even with a billion dollar fab to make such precise changes on a fabbed chips without destroying it.

 

[CTho9305]

Originally posted by: BrownTown

Originally posted by: firewolfsm

Originally posted by: BrownTown
I'm like 95% sure that they have fuses on the multiplier which are burnt in during fabrication for non-EE chips, it is physically impossible to unlock these chips.

Unless you can guess which of the hundreds of millions of transistors locked your chip

IT don't even matter if you know exactly where the problem is at, there is no possible way you could fix it. Just in theory you would have to start by debonding the chip from the packaging then etching off all the inter-level dielectric then etch away or deposit the correct vias to undo the fuses then regrow all the metal layers above etc..

IE: impossible even with a billion dollar fab to make such precise changes on a fabbed chips without destroying it.

No it's not. If you knew what to cut/reconnect, you could do it with a FIB. It could be as simple as disconnecting a transistor that pulls the "allow/disallow" signal(s) high or low so that it/they float(s) the way you want it/them to go.

 

[BrownTown]

Originally posted by: CTho9305

Originally posted by: BrownTown

Originally posted by: firewolfsm

Originally posted by: BrownTown
I'm like 95% sure that they have fuses on the multiplier which are burnt in during fabrication for non-EE chips, it is physically impossible to unlock these chips.

Unless you can guess which of the hundreds of millions of transistors locked your chip

IT don't even matter if you know exactly where the problem is at, there is no possible way you could fix it. Just in theory you would have to start by debonding the chip from the packaging then etching off all the inter-level dielectric then etch away or deposit the correct vias to undo the fuses then regrow all the metal layers above etc..

IE: impossible even with a billion dollar fab to make such precise changes on a fabbed chips without destroying it.

No it's not. If you knew what to cut/reconnect, you could do it with a FIB. It could be as simple as disconnecting a transistor that pulls the "allow/disallow" signal(s) high or low so that it/they float(s) the way you want it/them to go.

Well, I have no experience with all this, but I really just don't see it as feasible. Fuses are usually made by growing VIAs between metal layers using UV light after the chip is manufactured, or at least thats how the ones I have heard of work. So *maybe* if you had the layout and schematics you could imagine someway to go about it, but it isn't just some simple task of blowing away some transistor, you would have to cut through metal lines before you could ever get to that point and then when you were done you would have to regrow them exactly as they were before, and this is on a 65nm process. I just really doubt that those machines have that precision of etching/depositing so as to perform this task even if you know exactly where it needed to be done.

 

[CTho9305]

Originally posted by: BrownTown

Originally posted by: CTho9305

Originally posted by: BrownTown

Originally posted by: firewolfsm

Originally posted by: BrownTown
I'm like 95% sure that they have fuses on the multiplier which are burnt in during fabrication for non-EE chips, it is physically impossible to unlock these chips.

Unless you can guess which of the hundreds of millions of transistors locked your chip

IT don't even matter if you know exactly where the problem is at, there is no possible way you could fix it. Just in theory you would have to start by debonding the chip from the packaging then etching off all the inter-level dielectric then etch away or deposit the correct vias to undo the fuses then regrow all the metal layers above etc..

IE: impossible even with a billion dollar fab to make such precise changes on a fabbed chips without destroying it.

No it's not. If you knew what to cut/reconnect, you could do it with a FIB. It could be as simple as disconnecting a transistor that pulls the "allow/disallow" signal(s) high or low so that it/they float(s) the way you want it/them to go.

Well, I have no experience with all this, but I really just don't see it as feasible. Fuses are usually made by growing VIAs between metal layers using UV light after the chip is manufactured, or at least thats how the ones I have heard of work. So *maybe* if you had the layout and schematics you could imagine someway to go about it, but it isn't just some simple task of blowing away some transistor, you would have to cut through metal lines before you could ever get to that point and then when you were done you would have to regrow them exactly as they were before, and this is on a 65nm process. I just really doubt that those machines have that precision of etching/depositing so as to perform this task even if you know exactly where it needed to be done.

On a flip chip you're coming from the side with the transistors, and there's plenty of whitespace. FIBs are definitely precise enough. Editing logic is exactly the kind of thing they're used for in industry - sometimes, before taping out a fix for a bug, the fix is made with a FIB and tested. Sometimes a FIB doesn't work, or the node you need to hit isn't accessible, but it's a very useful tool.

 

[BrownTown]

meh, I guess maybe in theory, but in practice Intel ain't exactly giving out their layout files, so we would never know where to even begin. All academic really though since the FIB is a few orders of magnitude more expensive than the Extreme models, would be easier to jsut buy one, or if you wanna be cheap to steal one.

 

[dmens]

another option is if someone figures out how to unlock the test access port (TAP), which should allows ratio changes, among many, many other things.

just gotta figure out the binary sequence to feed the TDI pin. good luck

 

[covert24]

Originally posted by: dmens
another option is if someone figures out how to unlock the test access port (TAP), which should allows ratio changes, among many, many other things.

just gotta figure out the binary sequence to feed the TDI pin. good luck

a good weekend project for somebody

 

[BrownTown]

Originally posted by: dmens
another option is if someone figures out how to unlock the test access port (TAP), which should allows ratio changes, among many, many other things.

just gotta figure out the binary sequence to feed the TDI pin. good luck

well if you're gonna go about stealing the layout files for the entire chip you might as well get all the documentation too .

 

[HeXploiT]

Can't we just hire a Chinamen? They seem to be able to commit espionage with little trouble.
Well after Hillary is in office anyway.

 

[DisgruntledVirus]

Just buy the EE edition from Intel, or the Black edition from AMD and call it a day.

Or just do a FSB OC and call it a day

 

[HeXploiT]

Originally posted by: DisgruntledVirus
Just buy the EE edition from Intel, or the Black edition from AMD and call it a day.

Or just do a FSB OC and call it a day

I have an e6300 oc'ed to 3.2ghz via fsb. I have fried 2 680i motherboards with this same setup. My nb temp is 74. Who knows what temp the rest of the mb is. With an unlocked multiplier I might get 3.8 or 4.0ghz with very little excess heat.

 

[The Boston Dangler]

Originally posted by: Perry404

Originally posted by: DisgruntledVirus
Just buy the EE edition from Intel, or the Black edition from AMD and call it a day.

Or just do a FSB OC and call it a day

I have an e6300 oc'ed to 3.2ghz via fsb. I have fried 2 680i motherboards with this same setup. My nb temp is 74. Who knows what temp the rest of the mb is. With an unlocked multiplier I might get 3.8 or 4.0ghz with very little excess heat.

you're on your 3rd $200-250 motherboard, to squeeze more out of a $175 chip? you're either lying or very foolish.

 

[DisgruntledVirus]

Originally posted by: The Boston Dangler

Originally posted by: Perry404

Originally posted by: DisgruntledVirus
Just buy the EE edition from Intel, or the Black edition from AMD and call it a day.

Or just do a FSB OC and call it a day

I have an e6300 oc'ed to 3.2ghz via fsb. I have fried 2 680i motherboards with this same setup. My nb temp is 74. Who knows what temp the rest of the mb is. With an unlocked multiplier I might get 3.8 or 4.0ghz with very little excess heat.

you're on your 3rd $200-250 motherboard, to squeeze more out of a $175 chip? you're either lying or very foolish.

The Boston Dangler has a very valid point, if you have dropped that much on mobo's that you keep on frying why not instead spend the extra for an EE chip?

Or just not be as aggressive with the OC on that chip.

As for the "With an unlocked multiplier I might get 3.8 or 4.0ghz with very little excess heat", what are you smoking? If you know that much about OC'ing then you should realize that a chip hits a certain clock and for each incremental increase in clock you get a exponential increase in chip temps. On that chip I believe its around the 3.4 ghz range, but 3.8-4.0 ghz the cooling you would need in addition to the mobos burning up why not just drop the cash on an EE chip?

My guess is $500 extra in mobo's (2 burned up), $175 chip, means only $300 for the special cooling you could EASILY get an EE instead.

 

[HeXploiT]

Originally posted by: The Boston Dangler

Originally posted by: Perry404

Originally posted by: DisgruntledVirus
Just buy the EE edition from Intel, or the Black edition from AMD and call it a day.

Or just do a FSB OC and call it a day

I have an e6300 oc'ed to 3.2ghz via fsb. I have fried 2 680i motherboards with this same setup. My nb temp is 74. Who knows what temp the rest of the mb is. With an unlocked multiplier I might get 3.8 or 4.0ghz with very little excess heat.

you're on your 3rd $200-250 motherboard, to squeeze more out of a $175 chip? you're either lying or very foolish.

Hey fuck you. Call me a lier. In fact I was wise and bought my board from eVga who has a lifetime warranty and replaces the motherboard free no questions asked.