Signs of Intel segregating their processors into OCable and Stock

[Modelworks]

Originally posted by: myocardia


Which other brands are clock-locking their CPU's?

There are more things in heaven and earth........


Atmel, Motorola, TI, SST

All prove it can be done.
Motorola has chips that will boot fine, then once booted they do a self check, if the frequency is incorrect the chip goes into HLT mode.

TI has many processors that will not boot because the clock inputs to the chip are designed to detect improper frequencies down to .001 Mhz difference.

What I am saying is that the reason the current line of Intel cpu are not clock locked is not because they can't do it, its because they chose not to.

 

[bryanW1995]

Originally posted by: Foxery
I award FUDZilla 1/2 point for managing to write its first coherent article with no spelling mistakes.

+1

 

[Idontcare]

Originally posted by: bryanW1995

Originally posted by: Foxery
I award FUDZilla 1/2 point for managing to write its first coherent article with no spelling mistakes.

+1

It wasn't written by Fuad. It is credited to Lars-Göran Nilsson.

 

[myocardia]

Originally posted by: Modelworks
Atmel, Motorola, TI, SST

All are proprietary, AFAIK, quite unlike AMD & Intel.

Motorola has chips that will boot fine, then once booted they do a self check, if the frequency is incorrect the chip goes into HLT mode.

Yeah, that would work for Intel, now that you mention it.

Originally posted by: Idontcare
It wasn't written by Fuad. It is credited to Lars-Göran Nilsson.

I heard Fuad is off learning English, which makes sense, since he didn't seem to know any before.

 

[nyker96]

I doubt this will happen, if it does, Intel is giving AMD an entry into the enthusist market. AMD has had IMC type chips for years now without locking out the possibilities of overclocking. If Intel locks out the low end chips where some of us (myself included) uses to setup a OC system. They will most likely allow AMD to take over this segment entirely. Probably no biggie to Intel market sharewise but still why would they allow AMD into this segment just to force OCers to buy higher end chips which is just gonna turn away customers to his rivals. One thing Intel fears more is AMD get more market share, as for profit, they make enough already.

 

[Idontcare]

Originally posted by: nyker96
I doubt this will happen, if it does, Intel is giving AMD an entry into the enthusist market. AMD has had IMC type chips for years now without locking out the possibilities of overclocking. If Intel locks out the low end chips where some of us (myself included) uses to setup a OC system. They will most likely allow AMD to take over this segment entirely. Probably no biggie to Intel market sharewise but still why would they allow AMD into this segment just to force OCers to buy higher end chips which is just gonna turn away customers to his rivals. One thing Intel fears more is AMD get more market share, as for profit, they make enough already.

Perhaps for the same reason Microsoft is careful not to kill of Apple. (anti-trust insurance)

 

[hans007]

Originally posted by: Idontcare

Originally posted by: Foxery
Officially, Intel has never "supported" (or even condoned, as far as I know,) overclocking outside of Extreme Expense^H^H^H^H Edition products.

Unofficially, the 3rd-party motherboard and chipset makers always find a way, and I have faith that they will again. Phenom has an IMC and HyperTransport instead of FSB, and you can overclock Phenoms. Where there's a will, there's a way.

I would add to this by saying that Intel has always had the ability to stifle the overclocking community if they so desired by merely doing what it would take to enforce a policy on the mobo manufacturers such that no end-user available options in the BIOS were available for overclocking (i.e. no FSB options).

Even for non-Intel chipsets they merely need put a clause in their bus licensing agreements that stipulate 3rd party chipset providers would require same of mobo makers utilizing their chipsets as well.

Legally it would be a swift end to overclocking. For those who'd do BSEL mods or pin mods Intel could likewise simply require the BIOS force a CPUID reference check against a look-up table for allowed FSB settings and viola your system either boots at stock or doesn't boot at all. BIOS mods themselves could be blocked by further onboard reference checks that the mobo makers could be required to engineer into the mobo.

We know this is true because we have plenty examples of DELL and HP managing to accomplish most of this checklist of items of their own internal strategies.

That these requirements have never been pressed across the industry is more or less all the proof that I need to believe Intel has been just fine turning a blind'ish eye towards the communities un-approved activities thus far.

Will their policy change on the low-end? Here's the litmus test...if the rumors are true and their "hands are forced by coincidence" that the non-Bloomfield platform CPU's just so happen to not have external handles in the BIOS which allow overclocking and they do not release a software program as AMD did to allow enthusiasts to overclock from the OS then you can confidently assume the whole "overclocking only works on Bloomfield" was a strategy and not a coincidence.

If AMD can release a software program to do it then surely Intel could do so as well. This will be my litmus test whether a lack of overclocking on non-Bloomfield platforms was intentional or merely happenstance.

i figure someone willl find a way.

then again, i dont even overclock as it is anymore (i dont do anything that really need it) but i like the idea of that hobbyist scene as i once very much enjoyed it and it is what probably keeps me building my own isntead of buying a dell. just that i can mix and match etc.


oh well, hopefully amd has something to pull out of their ass.

 

[Sir Ali]

AMD is our hope then. They should come up with something competitives as their Athlon 64 line and intel would change their mind.

 

[bryanW1995]

Originally posted by: Idontcare

Originally posted by: bryanW1995

Originally posted by: Foxery
I award FUDZilla 1/2 point for managing to write its first coherent article with no spelling mistakes.

+1

It wasn't written by Fuad. It is credited to Lars-Göran Nilsson.

of course it wasn't written by fuad. if it had been then it would have been incoherent and had tons of spelling mistakes.

 

[bryanW1995]

Originally posted by: Sir Ali
AMD is our hope then. They should come up with something competitives as their Athlon 64 line and intel would change their mind.

amd doesn't have anything that will be competitive on the desktop until 2011 at the earliest. The best that amd can hope for it so MATCH PENRYN before mainstream nehalem comes out in 2009. If amd even does that then I would be shocked. Heck, I'll be impressed if amd gets to 3 ghz before lynnfield is out.

 

[Modelworks]

I don't see Intel clock locking the cpu unless its really worth the extra cost.
The people that overclock is such a small market compared to the total sales that I can't see them spending the resources to stop the process unless its either very easy to implement, or very cheap.

 

[Phynaz]

Originally posted by: zephyrprime

Originally posted by: Phynaz
I once read an article on why it would be impossile to lock a cpu clock.

This was around Pentium 1 days, I'm going to see if I can find it again.

It's only impossible if there's an external clock but if the cpu had and internal clock, then it would be possible.

Part of the article was why you can't put the clock on the chip and know what speed it's running at. I think the feedback loop still needs an external reference. Like I said, I need to find it again, but the internet is a big place.

Dmens or Ctho feel free to help me out here.

 

[Foxery]

Originally posted by: Phynaz
Part of the article was why you can't put the clock on the chip and know what speed it's running at. I think the feedback loop still needs an external reference. Like I said, I need to find it again, but the internet is a big place.

Can a single system even have more than one clock generator? I can only imagine them frequently falling out of synch or interfering in some manner. You would have to remove the motherboard's clock and put it on the CPU; redesigning a perfectly good system to do this instead would be a huge waste of time and resources.

I still say that any of these rumored hinderances are more likely to be side effects of the new architecture, rather than intentional design. And remember, they're nothing but rumors until we see silicon!

 

[Idontcare]

Originally posted by: Foxery
I still say that any of these rumored hinderances are more likely to be side effects of the new architecture, rather than intentional design. And remember, they're nothing but rumors until we see silicon!

I agree. At the same time I am all too ignorant of what material differences exist between AMD's implementation versus Intel's implementation such that Intel's would/could inherently preclude overclocking while AMD's did not.

 

[Modelworks]

Originally posted by: Foxery

Originally posted by: Phynaz
Part of the article was why you can't put the clock on the chip and know what speed it's running at. I think the feedback loop still needs an external reference. Like I said, I need to find it again, but the internet is a big place.

Can a single system even have more than one clock generator? I can only imagine them frequently falling out of synch or interfering in some manner. You would have to remove the motherboard's clock and put it on the CPU; redesigning a perfectly good system to do this instead would be a huge waste of time and resources.

I still say that any of these rumored hinderances are more likely to be side effects of the new architecture, rather than intentional design. And remember, they're nothing but rumors until we see silicon!

Yes a single system can have multitudes of clocks , its was done a lot in Risc systems like those Sun produced.


I don't have experience designing motherboards for x86 but I have designed embedded boards.
You can get embedded parts with the clock internally generated on the processor, but its at much slower speeds than the parts that use an external clock.

When you put the clock on the cpu you start having all sorts of issues to deal with for very little gain other than a lower part count on the board.

There is an old method for measuring the clock speed of a cpu that uses just one capacitor.
On power up you charge the capacitor through a resistor while at the same time executing a loop on the cpu of just 1+1. Once the capacitor is fully charged you stop the counting.
The value of the count has to fall within a certain range or you know that the clock speed is wrong and then can trigger whatever action you like.

 

[Phynaz]

The latest I'm hearing about this rumor is that Intel will remove overclocking capability from the chipset, not the cpu.

Now that makes much more sense.

 

[Viditor]

Originally posted by: Phynaz
The latest I'm hearing about this rumor is that Intel will remove overclocking capability from the chipset, not the cpu.

Now that makes much more sense.

But you have to wonder what that means as part of the chipset is being migrated to the CPU...

 

[Foxery]

Huh? Freely adjusting things isn't part of the official chipset spec anyway. 3rd party vendors add this functionality to what is intended to be a fixed/static system.

 

[myocardia]

Originally posted by: Viditor

Originally posted by: Phynaz
The latest I'm hearing about this rumor is that Intel will remove overclocking capability from the chipset, not the cpu.

Now that makes much more sense.

But you have to wonder what that means as part of the chipset is being migrated to the CPU...

Nothing is being moved except the memory controller, and as we already know, that has never stopped any of us from overclocking an AMD chip, with it's IMC. Either it's all BS, or Intel is using their sway in the industry to tell motherboard manufacturers that motherboards that allow overclocking won't be tolerated, and anyone making one that will overclock (for the lower-end chips) will lose their rights to make boards with Intel chipsets.

 

[Modelworks]

I seriously doubt they will ask / force manufacturers to stop allowing overclocking.
If the do it will hurt them more than help them.
Look at all the companies that depend on overclocking for an income.
All the specialty heatsink, cooling, memory, makers would lose out.

 

[Munky]

If this is true, then Intel has to compete in price and performance not only with stock AMD chips, but also oveclocked ones. In the long run, I think they would be shooting themselves in the foot by adopting such a strategy, and while it may lead them to better profit margins, it can't be that good for maintaining or increasing market share.