core i7-2700k is coming

[Mopetar]

Considering it's an unlocked part, is there any indication that the 2700K will be able to overclock higher? If not, I can't see any reason to prefer this over a 2600K.

 

[RussianSensation]

Why do people keep saying this? Yes, the mobo will be slightly more $$, but the RAM is the same price. In fact, it is cheaper to buy four 2GB sticks (quad channel) than it is to buy two 4GB sticks (dual channel). So that arguement is not valid. Tripple channel was more expensive only because you have to buy special kits of 3. That is not going to be the case with quad channel.

Because it's true.

1) Who is going to spend $300-350 on a motherboard only to fill all of its banks with 4x 2GB sticks? You are pretty much getting 16GBs of Ram at minimum with LGA2011, but if you need 16GBs of Ram, then surely you are running workstation apps that need 6 cores, not a 4 core CPU. Suddenly LGA Platform overall cost is > $950 ($580 CPU + $300 motherboard + $100 16GB of Ram). That's not competing with BD or 2600k/2700k platform whatsoever.

2) If you only buy 2x 4GB sticks, then you aren't getting quad channel, so why are you paying a premium for LGA2011 then?

3) 4x2gb is NOT cheaper than 2x4GB:

http://www.newegg.com/Product/Produc...82E16820145343

vs.

http://www.newegg.com/Product/Produc...82E16820233144

4) X79 1st revisions may not even have PCIe 2.0. So again, you are paying even more $$$ for X79. Why? X79 is shaping up to be a disappointment. I mean if you really need 6 cores, it makes sense to wait since 6 core will drop to < $600. But 4 core models on X79 make absolutely no sense.

5) Most LGA1366 users aren't going to be making the same mistake again. Also overclocking BLKC at 100/133/166 increments may not allow you to extract maximum overclocking since it's not a gradual overclocking progression like FSB or QPI was on a +1 increment basis.

Not only that, you can now get Asrock Extreme4 Gen 3 with PCIe 3.0 for $155. We also know that memory bandwidth is not a factor for quad-core SB processors. With the argument behind a quad-core X79 platform is non-existent unless you like wasting $$ for E-Peen that you own the bottom feeder X79 setup.

 

[blackened23]

how many times do I have to state 22nm for you to actually read it and understand ?

todays 2500-2600 are on 32nm

You are seriously borderline (radio edit) if you think intel is going to make a 4.4ghz 2700k as 22nm and have it similarly priced as 2600k. Here's how CPU's are binned: A newly fabbed CPU is connected to a special "tester" machine. This machine controls all of the pins to the CPU, and specially-written test programs are played onto the pins (with expected results). The clock and voltage are increased until the point of failure, and this determines which bin it makes.

The 2700k is being binned the same as the 2500 and 2600. Simply upping the clock isn't a trivial matter. While it may overclock to 4.4 easily, its not as easy to have a good yield for a 4.4ghz part. Since the 2700 will be binned the same as 2500/2600, its a near certainty that they A) not going to be 22nm B) be clocked any higher than a 100-200 mhz increment. Anything more than a 100mhz increment would likely have poor yields/profitability. This is speculation on my part, but i'm almost certain that 2700k will not exceed 4ghz turbo frequency.

http://www.anandtech.com/show/4757/intel-leaks-i72700k

Speculation from those in the know is that it will actually be a 3.9 ghz turbo, 3.5-3.6 base. Feel free to argue with every other website speculating the same thing with your broken english.

 

[Rvenger]

You are seriously borderline (radio edit) if you think intel is going to make a 4.4ghz 2700k as 22nm and have it similarly priced as 2600k. Here's how CPU's are binned: A newly fabbed CPU is connected to a special "tester" machine. This machine controls all of the pins to the CPU, and specially-written test programs are played onto the pins (with expected results). The clock and voltage are increased until the point of failure, and this determines which bin it makes.

The 2700k is being binned the same as the 2500 and 2600. Simply upping the clock isn't a trivial matter. While it may overclock to 4.4 easily, its not as easy to have a good yield for a 4.4ghz part. Since the 2700 will be binned the same as 2500/2600, its a near certainty that they A) not going to be 22nm B) be clocked any higher than a 100-200 mhz increment. Anything more than a 100mhz increment would likely have poor yields/profitability. This is speculation on my part, but i'm almost certain that 2700k will not exceed 4ghz turbo frequency.

http://www.anandtech.com/show/4757/intel-leaks-i72700k

Speculation from those in the know is that it will actually be a 3.9 ghz turbo, 3.5-3.6 base. Feel free to argue with every other website speculating the same thing with your broken english.

Exactly what I was thinking since the trends have been:

2500k - 3.3ghz/3.7ghz Turbo
2600k - 3.4ghz/3.8ghz Turbo

2700k - 3.5ghz/3.9ghz Turbo

This is not AMD where they do things that make no sense. ie- have a quad core Stock 4ghz with a 100mhz turbo and an 8100 with a 800mhz turboD:

 

[Spikesoldier]

only thing that can be exciting is if its a new stepping.

if not, yawn.

 

[gevorg]

Exactly what I was thinking since the trends have been:

2500k - 3.3ghz/3.7ghz Turbo
2600k - 3.4ghz/3.8ghz Turbo

2700k - 3.5ghz/3.9ghz Turbo

This is not AMD where they do things that make no sense. ie- have a quad core Stock 4ghz with a 100mhz turbo and an 8100 with a 800mhz turboD:

i7-3200K @ 4.0Ghz would be pretty awesome for marketing and stock benchmarks, even if its the same as current 2600K but at higher stock settings. i7-2700K on the other hand is just meh.

 

[Soccerman06]

only thing that can be exciting is if its a new stepping.

if not, yawn.

My exact thoughts. Would be interesting to see if a new stepping would let the average unit hit 5ghz

 

[Skott]

Yeah, either way I'm not excited. Its a little interesting but nothing else. Kinda ho hum I guess.

 

[Arachnotronic]

i7 2600K is crazy fast. i7 2700K for the same price will also be crazy fast. I'm not seeing a downside here. Want more performance now? Go x58. Want more performance in a few months? Go x79.

 

[dac7nco]

i7 2600K is crazy fast. i7 2700K for the same price will also be crazy fast. I'm not seeing a downside here. Want more performance now? Go x58.

X58 is slower, buddy. Westmere is faster in a few things, but that's very few. X58 also gulps juice like a teamster, and runs a lot hotter to boot.

Daimon

 

[BTRY B 529th FA BN]

I'd imagine all you are paying for is a higher default speed. What good is that on a K chip? What more does that offer aside from an i7 2700 chip, IF they put out one of those.

 

[Arachnotronic]

X58 is slower, buddy. Westmere is faster in a few things, but that's very few. X58 also gulps juice like a teamster, and runs a lot hotter to boot.

Daimon

i7 970/980/990x slower than 2600K?

 

[dac7nco]

i7 970/980/990x slower than 2600K?

In everything except for encoding and rendering, yes. Don't take my word for it, read a review or two.

Daimon.

 

[blackened23]

X58 is slower, buddy. Westmere is faster in a few things, but that's very few. X58 also gulps juice like a teamster, and runs a lot hotter to boot.

Daimon

I'm not sure what you're talking about. Look at the featureset the x58 offers - it offers better features than the p67/z68, and is the only chipset to support x16/x16 sli/crossfire. All p67/z68 variants only do a max of x8/x8 unless the nf200 nvidia chip is used, which scales horribly and is pricy.

The x58 chipset *is* better, however it supports a dated cpu architecture. I think what you meant to say is, the CPU's supported by x58 are dated.....hence why x79 with dual x16 and sb-e are being developed as a successor.

 

[blackened23]

i7 970/980/990x slower than 2600K?

slower in some applications, faster in others. For gaming 2600k is better. With that in mind, games are being developed for old technology due to the games industry being driven by AAA titles that are cross platform - no games are truly tapping the potential of modern pc's....if games were developed with modern hardware in mind the extreme processors would likely be far better. Anyway, for true productivity applications, 980x/990x are better. Overall, the 980x/990x are better. All of the above CPU's overclock well.

I really don't think intel intended SB to trump extreme processors (for certain applications), the SB is just extremely strong. This is all being rectified with SB-e, which will be the successor to current 980x/990x cpu's and will trump the 2500/2600k.

 

[Idontcare]

i7 970/980/990x slower than 2600K?

Anandtech bench results show they trade back and forth at stock clocks, and you'll be hard-pressed to get a gulftown up to an OC'ed clockspeed that SB won't eclipse if given equal effort with the cooling setup and expertise.

In my opinion the 2600K vs. 980/990X case is exactly the argument that AMD is bringing to the table with their Zambezi vs. 2600K lineup...the argument being that the lower single-threaded IPC (be it Zambezi or Gulftown) is irrelevant because they quantity trumps quality and they make up for the gap by adding more cores.

Only numbers can annihilate.

So whether you find yourself arguing for the imminent superiority of Zambezi over SB, or the existence of a superiority of Gulftown over Sandy Bridge at present, you are really basing your argument on the same footing...more cores w/lower-IPC wins the day provided the clockspeed is just enough.

And its not an unfamiliar model. If you look at the performance capabilities of an individual NAND chip that goes into your SSD you'll be appalled at the write latency and overall bandwidth numbers a single chip can support. What makes SSD's deliver out-of-this-world performance is a combination of the fact they gang up 8, 16, and even 32 of these individual chips - ala a crude multi-core style - and the fact that their nearest technological competitor hearkens from the 1950's and is an electromechanical device the likes of which Robby the Robot would have inside.

Many hands make for light work, Gulftown is a beast, the only thing we lower-budgeted enthusiasts don't like about it is the pricetag and the planned obsolescence of the platform/socket.

 

[Diogenes2]

Anandtech bench results show they trade back and forth at stock clocks, and you'll be hard-pressed to get a gulftown up to an OC'ed clockspeed that SB won't eclipse if given equal effort with the cooling setup and expertise.
.....

I do video encoding with Sony Vegas Pro.. Have to run my SB at 4.9 to match my 980X at 4.2 .. And with less vcore ..
The 980X uses 1.3 while it takes 1.45 to get the SB to 4.9


While both overclocks require very little tweaking, and temps are acceptable with a CM 212+, the 980X runs cooler - SB mid 70's, the 980X mid to high 60's..

Off course the price/performance is another matter ..

 

[gevorg]

i7 970/980/990x slower than 2600K?

In everything except for encoding and rendering, yes. Don't take my word for it, read a review or two.

Daimon.

Once you overclock 2600K it owns in most benchmarks, especially gaming:

http://www.xbitlabs.com/articles/cpu/display/core-i7-2600k-990x.html

 

[RussianSensation]

i7 970/980/990x slower than 2600K?

Pages 9-12. Oh Gevorg beat me to it!

slower in some applications, faster in others. For gaming 2600k is better. Overall, the 980x/990x are better.

Not just in gaming. Here is another review. Really, there should never be any debate when comparing a $1000 CPU to a $200-300. The fact that we have a debate means that 980/990X are not better at all. You are paying $1000 for a CPU that loses in most scenarios. A $1000 CPU should win every single benchmark, no questions asked, and yet it struggles to beat a $225 2500k in half the benchmarks (including a lot of common office apps). Anyway, this debate won't exist once SB-E 6 cores are launched.

 

[Idontcare]

A $1000 CPU should win every single benchmark, no questions asked.

That's the nut of it right there.

A $1000 "extreme" processor should not be taken to the mat in anything by a $300 processor.

Add to this the fact that if you buy that $1000 chip today you are buying a soon-dead platform too and you've got yourself a real

 

[greenhawk]

They should release some more i5 'k' variants as well. It makes me wonder if they are going to also possibly phase out the 2500k and just leave 2700k and 2800k.

intel might phase out the 2500k, but my money would be on a 2550 that is slightly higher clock, higher turbo range, but no "k" option.

but getting rid of 2500 all together is not going to happen. Leaves too big of a price jump from the 2400 to the 2600.

 

[greenhawk]

I only wished there 1155 ivys came out since I need a new laptop and hp has 30% off coupons since they are getting out of the pc market.

I would be surprised if HP released a Ivy Bridge laptop if they are looking to get out of the market. Doing any form of R&D for a new laptop on the way out the door does not strike me as smart business. But then the markets HP has been getting out of (and how) at a loss I find strange anyway.

 

[greenhawk]

As to the original part number, I wonder it is more to do with Ivy Bridge (ie: setting up part numbers in advance) than just a re-adjustment of the 2600 range.

but then marketing might want a bigger gap in names to warrent the price difference (base cpu cost).

 

[gevorg]

Maybe Intel has some insider info on Bulldozer performance and need to nudge performance of 2600K price bracket a bit. But thats only if BD is competitive.

/tinfoilhat

 

[podspi]

I would be surprised if HP released a Ivy Bridge laptop if they are looking to get out of the market. Doing any form of R&D for a new laptop on the way out the door does not strike me as smart business. But then the markets HP has been getting out of (and how) at a loss I find strange anyway.

They want to spin off their PC division, not wind it down.

If they stop all R&D then the division won't be worth much by the time somebody gets around to buying it, or IPO, etc.

From a business prospective what they're doing makes sense, just as it did for IBM. Get out of the super competitive and low margin markets, and focus on services. I'll be sad to see them go, though.