Unlocked Clarkdales and Lynnfields coming our way?

[Massman]

Intel Corp. plans to release inexpensive central processing units (CPUs) with unlocked multipliers designed for overclockers in an attempt to offer appealing products for enthusiasts, who do not want to spend to much on personal computers.

Historically, neither of the two leading suppliers of central processing units accepter overclocking since this could damage their chips in many cases. However, when both Advanced Micro Devices and Intel started to sell processors designed for extreme performance enthusiasts, they had to admit that such chips are often used at clock-speeds that by far exceed the official ones. As a result, both AMD and Intel have been shipping their chips for overclockers – Extreme Edition, Black Edition, FX-series – with unlocked multiplier for about seven years now. AMD started to offer its Black Edition chips with unlocked multiplier and moderate price back in 2007 and it looks like Intel plant to follow its smaller rival this year.

Steve Peterson, head of marketing of Intel's chipset division, said in an interview with Tweakers.net web-site that Intel was considering new affordable models with unlocked multiplier. Even thouth Mr. Peterson did not elaborate a lot, he did reveal that the chips would be designed for LGA1156 form-factor and will belong to Lynnfield and Clarkdale families of Core i-branded chips.

It will be interesting to see how unlocked multiplier and overclocking will live on with Intel's Turbo Boost technology that automatically overclocks select processing cores when they are truly needed and disables the unused ones. Turbo Boost disables itself in case of overclocking of modern Extreme Edition microprocessors, so, end-users will have to count only on themselves if they need massive speed in single-threaded or dual-threaded applications and they overclock an inexpensive quad-core chip.

Back in August ’09 the world’s largest chipmaker already tried to sell its Pentium 6500K processor with unlocked multiplier in China as a pilot program. The company said back then that based on the demand towards the product it would make the decision about selling similar CPUs in other regions.

“This is a PRC-only pilot for now. At the end of the pilot, Intel will determine what, if any other similar product should be offered in other geographies,” said Daniel Snyder, a spokesperson for Intel.

~ Xbitlabs (http://www.xbitlabs.com/news/cpu/display/20100303151423_Intel_Plans_to_Offer_Inexpensive_Mi croprocessors_with_Unlocked_Multiplier.html)

[Massman]

I'm having the biggest grin on my face since Conroe performance results made it to the internet ... this will be amazing!

[jabski]

sounds good :)

[bazx]

i thought they aready did unlock for 6500k

this is for sure the way to go

[Linuxfan]

Sounds good to me!

[Gamer]

:banana: Intel :celebration:

[Eeky NoX]

OMG 3rd news of the day to make me jump ...better than Xmas though! :D

Well now I see how I'm stuck with my i5 750 multi, it's the best thing I heard today \o/ that's the way, sounds SO GOOD yeapie ^^

[CoolnQuiet]

Good news indeed! With affordable unlocked CPUs freely available more people will overclock. This could help to get overclocking recognized as an e Sport.

Any news on pricing? I'd pay a small premium for unlocked CPUs.

[scooter.jay]

Should make things fun for all levels of cpu:)

[Massman]

Just thinking about it ... does unlocked means "cpu multiplier", "uncore multiplier" or "memory multiplier" unclocked? I mean ... if you want either Lynnfield or Clarkdale to match GT performance, you'll need more than just raw CPU speed.

[Antinomy]

I'll add to your thoughts - the idea is that the overclocking capabilities shouldn't be cut off.
The E6500K that I have in my system now (and as I know, you've been testing too) has too weak o/cing. We've checked five CPUs and only one took 4GHz at 1.3-1.35. This is rubbish being compared to E8000. Add the cache being cut and you'll see a poor scene.

In fact the FSB wall of this CPU is over 500MHz as I remember so it doesn't need the multi using air and water cooling. Maybe only an SS and LN2 can show an improvement but I doubt. So the E6500K is a failure. If Intel won't be so greed then this will make sense.

But they are greed :D

[Eeky NoX]

I just have to stop you a bit my friend on that purpose :D
I found that chip very smooth to OC on my Commando, and no wall @500 'cus of this test conclusion ;)

...very large fsb scaling, massive headroom ^^
http://www.overclocking-tv.com/forum/index.php?topic=215.msg1771#new

So don't matter what the mobo is able to rise fsb when the multi is unlocked ...even old nforce can handle it with relativ high freq
Then the failure isn't so commercial when you think about the massive amount of PCs which can handle that kind of chip ;)

But the cut of cache and the juice it needs are some bad points I agree with you ;) Failure for benching efficiency that's right but not for fun in fact, I like to play Amd with a blue chip lol ...and now I'd like so much to do the same with Lynfields and Clarkies

I've never touched any Xtrem chip in my life and it might be forever regarding pricing :/
...I use to play with SS and I'm affraid with the way it goes on my Lynfield as I said ...I was expecting more headroom :(

[Antinomy]

So don't matter what the mobo is able to rise fsb when the multi is unlocked ...even old nforce can handle it with relativ high freq O.K. and again - your max frequency (screen only, not benchable) is what - 4560? Wolfdale-6M can do this at 1.45-1.5 or even lower. And they are benchable. If you take a usual E6500 this will require only 418FSB - even a super old 975X (the first to support Core2) could handle this. But the benchable frequency is about 4300 on air - below 400.
The only case that can't take this is the 945GC maybe :D
If you have money for an SS to push it further you certainly do have some for a crappy P965 or P35 MoBo - and they do 500MHz easily.
So my statement is that these CPU's don't need free multi no matter the case.

Or maybe a massive amount of PC's have SS or cascade? I don't think that they couldn't afford a E7300-7500 which goes better. Even a good E5200 can outstand E6500K if it's lucky. On the contrary - the AMD 720BE is a nice one. Even if we don't take the unlockable core.

If we had a E7000 or 8000 with an unlocked multi - then yes. Or even better - a quad core CPU.
And yes, my Nforce 780 board took about 520 (didn't test more - after changing thermal greese it looks like it died) - far more than enough for this stone. On it's default multiplier it's not FSB limited on all MoBo's except 945GC.

I hope that Lynnfield will do better - it was Intel's first try after all. And I've never had an XE chip either.

[Massman]

If we had a E7000 or 8000 with an unlocked multi - then yes. Or even better - a quad core CPU.

Ehr ...

http://hwbot.org/hardware/processor/core_2_qx9650_3.00ghz
http://hwbot.org/hardware/processor/core_2_qx9770_3.2ghz

[Antinomy]

I've never touched any Xtrem chip in my life and it might be forever regarding pricing :/
Massman, the words weren't about CPU's that only some can afford. The topic is about only mainstream processor with unlocked multipliers :)
I don't have a spare 1000$ and if I did I would spend better the half on old school stuff or at least on a less expensive CPU. And the rest on my girlfriend. I not zealous enough for this yet :D

[Massman]

Then it's the typical "I want everything for nothing"-argument :D.

FYI, the 6500K was probably not to promote the overclockability of the architecture, but merely to check if there's any interest for this kind of product. The fact that many tried to get one and tried to clock them as hard as possible. It's better to do this with a low-cost architecture than high-cost.

In addition, unlike the Clarkdale/Lynnfield, the Wolfdale architecture was running at the end of its capabilities already with the E8600.

[Antinomy]

Then it's the typical "I want everything for nothing"-argument :D.

FYI, the 6500K was probably not to promote the overclockability of the architecture, but merely to check if there's any interest for this kind of product. The fact that many tried to get one and tried to clock them as hard as possible.Maybe :D

Yeah, I got one because of the Asrock 4CoreDual - it's pretty FSB-limited. And because I have China in only 200-250km from me :)
The reason to push these high is that they cost much lower than extreme edition series. Perhaps, you're right about 8600 though good batches of E8700 seem to give a bit more fun. I've found were to buy one of these.
As I said - we'll see what will happen. I'm hoping it'll be a nice one :)

[Massman]

Perhaps, you're right about 8600 though good batches of E8700 seem to give a bit more fun. I've found were to buy one of these.

The statistical significance of Knopflerbruce's E8700 is, well, non-existant. Okay, it's quite a good chip since it's the only one that is able to run 1M at 6.7G, but apart from that there's no data yet. How does it perform in 01? And how in other benchmarks?

Still a long way before we can say that E8700 is indeed better than E8600/.

[knopflerbruce]

The statistical significance of Knopflerbruce's E8700 is, well, non-existant. Okay, it's quite a good chip since it's the only one that is able to run 1M at 6.7G, but apart from that there's no data yet. How does it perform in 01? And how in other benchmarks?

Still a long way before we can say that E8700 is indeed better than E8600/.

But we can't say they're worse either - tbh it seems as if you pick a random e8600 and a random e8700 you'll get roughly the same results.:)

[Massman]

Thus proving the point that the Wolfdale architecture was already put to its end with the E8600 and there was no need for higher binned products ... :)

[knopflerbruce]

Thus proving the point that the Wolfdale architecture was already put to its end with the E8600 and there was no need for higher binned products ... :)

Someone once told me that there COULD be FSB limitations on good E8600's on LN2, in those cases an extra .5 multi would help.

The problem about drawing conclusions about E8700 is that there are only two batches (three if you include the ES batch that hasn't been tested by anyone yet) out there, if they made alot more I'm sure there would be somewhat better scores with these chips. But some things seem pretty clear:

1: The bad chips are not FSB limited, so the extra half multiplier doesn't help there.

2: A gazillion E8600 have been tested, and so few E8700's. Still the only one that completed an 1m run at 6.7+ was an E8700. The probability that "something" made the good E8700's better than the good E8600's is pretty high. I mean... there must've been tested thousands of E8600's on LN2, and like... 30 E8700's? To me it seems pretty clear that there are E8600's out there that could be potential 6.7GHz chips with a 10.5 multi, so the existing chips were not always maxed out with the 10 multi.

[Antinomy]

I fully agree with you, Massman. Though there are results of Turrican on E8700 for example with 8800GT. 6GHz under cascade is nice.
E8700 can't bring significant difference - because you can select a E8600 out of ten-twenty and the E8700 out of two, perhaps.
At top speed E8600 limits in motherboard capabilities - so in this case E8700 will be a bit more useful. Nothing significant, I repeat.
So I hope we won't see a situation when an unlocked multiplier is useless because of the price/performance level or because the unlocked model is so cut so it makes it useless against a bit more pricy ones.

[Brian y.]

There will be a 665k dual core(clarkdale) and a 875k that will be quad core.