System builders are not happy with Ivy Bridge

bulldog147 · 1 May 2012 at 00:51

Stock speed is a bit boring to be honest @ 3.50GHz (3.90 with Turbo) They should get 4.00GHz stock by now (4.40 with Turbo)

sunama · 1 May 2012 at 02:26

VaderDSL said:
Indeed, they're starting to push the current tech to the edge. There will be a point (we may have started to see evidence on 22nm) where laws of physics don't allow current tech to progress past.

I agree that laws of physics are getting ever closer.
However, to say that the 22nm process has started to reach those laws is incorrect. The same could be said for the Prescott CPUs, which also ran hot but used the far larger 90nm process. That was many years ago and since then Intel have released CPUs which are probably around 100 times faster.

So, while I think the laws of physics are getting closer and closer, I do not think we are anywhere near the limit. There is a reason why Intel has a such a huge R+D budget - its so that they can push their technology ever closer to the laws of physics. And guess what, in the last 4 decades they have succeeded every time.

Krono5 · 1 May 2012 at 03:01

The Halk said:
You are right according to the law, however it would be pointless trying to actually enforce that with Intel.

First of all they've no way to prove what voltage and clock speed you ran it at, and secondly even if they could you'd need to take them to court which would wipe out any financial gain.

No need to enforce it with intel, it would usually be the shop you buy it from that you deal with. They've no way to prove that you even overclocked it. Court isn't necessary from what i've heard as just a mention of trading standards etc gets it sorted.

Personally I'd never try to RMA something that I pushed to far, but I suppose it's how much of a douche you want to be about it as the consumer has a lot of power in this instance when they know how to excersise their rights.

Yaayuh! · 1 May 2012 at 07:07

Krono5 said:
It doesn't according to Intel. However according to UK law; specifically the sales of goods act 1979, EU directive 1999/44/EC and the actual warranty it does.

It's advertised as a K model so can be unlocked for overclocking, the specifications list 1.52v as the max volts so a reasonable person would say overclocking a K model chip as long as you stay within specifications like TJMax/Tmax/Vcore max is normal use and any failure would be because of faulty manufacturing as opposed to consumer misuse within the reasonable lifespan of the processor.

The key point is that it's advertised as unlockable multiplier upto 63*, so it is assumed that is meant to be used as a feature of the CPU.

Hahaha! What?! No. That's not how it works. That's like saying your car speedo goes upto 200mph but your car only does 150mph after modifications and you moan to Ford about it.

At no point have Intel said you can reach any overclock or advised that you should. You can't just go on what you expect to get out of it based on how the chip allows you to configure it. Intel (a long with me) will laugh in your face.
Intel have sold the chip as is, no guarantee of overclock, there's nothing wrong with it.
Many chips back in the day had completely unlocked chips, and I didn't get 6Ghz out them, maybe I should complain?
You sir, crack me up.

Krono5 · 1 May 2012 at 07:39

Rossi~ said:
Hahaha! What?! No. That's not how it works. That's like saying your car speedo goes upto 200mph but your car only does 150mph after modifications and you moan to Ford about it.

At no point have Intel said you can reach any overclock or advised that you should. You can't just go on what you expect to get out of it based on how the chip allows you to configure it. Intel (a long with me) will laugh in your face.
Intel have sold the chip as is, no guarantee of overclock, there's nothing wrong with it.
Many chips back in the day had completely unlocked chips, and I didn't get 6Ghz out them, maybe I should complain?
You sir, crack me up.

You sir need to learn to read instead of going off on a tangent perpendicular to anything I have said, of which at no point in this thread has congrugated around your imagined thesis of anyone complaining about unattainable overclocks.

P.S. I object to your tone, particularly considering the stupidity of your post. May I suggest in future when you throw disparaging remarks around you ensure you are correct otherwise it will again make you look mightily foolish.

Cythx · 1 May 2012 at 13:04

Krono5 said:
You sir need to learn to read instead of going off on a tangent perpendicular to anything I have said, of which at no point in this thread has congrugated around your imagined thesis of anyone complaining about unattainable overclocks.

P.S. I object to your tone, particularly considering the stupidity of your post. May I suggest in future when you throw disparaging remarks around you ensure you are correct otherwise it will again make you look mightily foolish.

Congregate

555BUK · 1 May 2012 at 14:14

IB is a tick. The big gains come after a tock (once the process has been refined and new architecture implemented). Swtching to 22nm is more about providing Intel with bigger margins, rather than giving consumers massive performance gains. The carrots for consumers are slightly lower power consumption, enhamced HD4000 graphics, and PCI-E 3 support (although the HD4000 graphics and PCI-E 3 kind of cancel eachother's relevance out).

Krono5 · 1 May 2012 at 14:25

Cythx said:
Congregate

Cythx said:
Probably more than 5% performance increase and not sure how many people expected impaired overclockability.

Also, if 1 year is truly such a short time to expect a good improvement, as you say, people wouldn't be this upset. But I'm guessing they've come to expect more than 5% performance improvement in over a years time, so I'd guess that's part of what their disappointed about.

They're

paullewis323 · 2 May 2012 at 11:48

Guys

Surely as you go thinner on circuits (silicone) you will always be prone to higher temperatures due to heat dissipation?

Lew

sunama · 2 May 2012 at 12:10

This is generally true. But then it is up to Intel (or AMD), to ensure that there is adequate cooling to prevent super high temps.

Apparently, Intel have had a few problems with the cooling (ie, they used a TIM instead of solder), which has meant that the heat transfer isn't as efficient as it otherwise might've been.

sunama · 2 May 2012 at 12:11

Cythx said:
But I'm guessing they've come to expect more than 5% performance improvement in over a years time, so I'd guess that's part of what their disappointed about.

This is exactly my point.

5% improvement in 1 year is disappointing.

G-Sharp · 2 May 2012 at 13:18

sunama said:
This is exactly my point.

5% improvement in 1 year is disappointing.

Is this at the same clock, or both at their max stable clock?

sunama · 2 May 2012 at 13:30

Both.

Remember, at their stock clock speed, Intel have been sneaky - they have raised the clock speed by 100mhz to give their new CPU the edge, over their old CPU.

Personally, I'm only interested in maximum possible stable overclock speed and in real terms, the improvement over SB is minimal (almost negligible, assuming that SB can overclock 100mhz or so higher than a IB). For this reason, I have put in an order, 13 hours ago, for a Z77 motherboard and a 2600K. The extra cost of an IB just isn't worth it (for me).

mmj_uk · 2 May 2012 at 13:45

IB is more of a GPU upgrade than a CPU upgrade, for those people buying PC's with no discrete graphics card fitted it is a significant improvement.

People wanting a fast CPU with good clocking ability should stick to SB for now, unless you have good cooling.

sunama · 2 May 2012 at 14:01

I am probably going to be using the on-die GPU, to power 2 displays.
However, as I shall not be gaming, I am not interested in paying more money for the HD4000, which is on the IB. If you are gaming, then really, paying extra for the HD4000 makes little sense. It would be better for you to buy a £40 video card, which will outperform the HD4000.

Whichever way you look at it, it makes no sense to buy the IB, over SB, unless the pricing is similar and you intend to run it at stock speeds.

For laptops, Ivy Bridge all the way.

I bought the 2600k, rather than pay £50 extra for the IB.

muon · 2 May 2012 at 14:51

I also bought a retail 2600k@£180 last week for my z77 board instead of waiting for IB. I am glad I did as my Antec Kuhler 620 would have been woefully inadequate.

If a better revision comes out e.g. G0 for Kentsfield and E0 for Yorkfield, then I'll think about IB. However SB never saw a second revision, so possible IB doesn't get one.

Nightglow · 2 May 2012 at 15:06

Disappointed with IB, think I will stick with a 2600K or 2700K until Haswell.

With IB not delivering what was expected, plus heat issues, I wonder whether AMD has something waiting to match or surpass IB.

prescott28 · 2 May 2012 at 15:24

My Ivy just arrived along with the other upgrade bits - moving up from Q6600. I hope I made e a good choice but, frankly, a 4.5GHz over-clock will be fine.

sunama · 2 May 2012 at 15:32

anyon said:
I also bought a retail 2600k@£180 last week for my z77 board instead of waiting for IB. I am glad I did as my Antec Kuhler 620 would have been woefully inadequate.

Have you overclocked the 2600K yet?

sunama · 2 May 2012 at 15:34

prescott28 said:
My Ivy just arrived along with the other upgrade bits - moving up from Q6600. I hope I made e a good choice but, frankly, a 4.5GHz over-clock will be fine.

I only hope you that you don't end up with another "Pres - hott".

Let us know how the overclock goes - even now, there is very little information on retail overclocked IB CPUs.