I guess we disagree, less power consumption (which is what Broadwell brings) will provide MORE headroom (along with other efficiencies).

 

My point is there has already been a case of reduced power consumption brought about by a process node shrink which did the *OPPOSITE* of what you claim. Your claim flies in the face of the established history because we have hit a thermal/voltage wall.

 

I guess we disagree again ... FCC influences motherboard designs please read: http://transition.fc...62/oet62rev.pdf Sections 15.107, 15.109, 15.101a and more.

 

EMI produced at 30 meters by a CPU in a shielded case is probably less than cosmic background noise. 150W is not a hard cap set by any regulatory body, it is a number chosen by Intel which suits their design and manufacturing purposes. You'll note the number has changed over the years as well, yet that document you've cited was last revised in 1996. Therefore, the FCC has nothing to do with CPU TDPs.

 

Unless you performed these "tests" in a controlled environment and went over the results with a multitude of EEs, I wouldn't draw too much of a conclusion. In my own experience, reducing voltage along with other EFI settings has resulted in higher achievable clock rates upto 5Ghz and a very stable platform. But to each his or her own, glad you believe in your methods and they work for you.

 

You wouldn't draw too much of a conclusion by repeating a test numerous times within a fixed range of parameters and achieving similar positive results that scale within the expected range in accordance with the defined parameters? How then would you test? What would you take away from the results? As for undervolting producing higher clockspeeds - you're really off in looney land on this one, ever hear of a shmoo plot? Go ask one of your engineer buddies about it.

you're really off in looney land on this one, ever hear of a shmoo plot?

 

Your derogatory comments still don't improve your position and understanding but I realize that is a common form of "intellectual assault" on internet forums where some individuals have a very strong conviction. Your logic fails in many areas and you also introduce assumptions and concepts that weren't suggested or presented.

 

I can only recommend you read up on more recent FCC documents if you don't value one from 1996 (they have not changed too much for computing devices with some exceptions to mobile devices, but that's not pertinent to this topic) - they are all available online.

 

No I wouldn't draw much from your conclusions unless you are indeed certified and have support from independent parties confirming your conclusions and they are also certified. I'm not making any definitive conclusion but you were -- specifically

flux-less solder vs. thermal paste has been demonstrated to be responsible for the loss of several hundred MHz ... Intel's ridiculous decision to use thermal paste instead of flux-less solder

 

If I were to be presented with such testing it would need to be with EEs and a certified body, not

I know, I've done it on 4 different Ivy Bridge chips and have posted the results in this very forum.

 

Anyway, I can see where this thread is going to end up, you seem unable to allow someone to politely disagree with your position and will continue heckling comments until either the thread is locked or people just stop responding.

Your derogatory comments still don't improve your position and understanding but I realize that is a common form of "intellectual assault" on internet forums where some individuals have a very strong conviction. Your logic fails in many areas and you also introduce assumptions and concepts that weren't suggested or presented.

 

I can only recommend you read up on more recent FCC documents if you don't value one from 1996 (they have not changed too much for computing devices with some exceptions to mobile devices, but that's not pertinent to this topic) - they are all available online.

 

No I wouldn't draw much from your conclusions unless you are indeed certified and have support from independent parties confirming your conclusions and they are also certified. I'm not making any definitive conclusion but you were -- specifically

 

 

If I were to be presented with such testing it would need to be with EEs and a certified body, not

 

 

Anyway, I can see where this thread is going to end up, you seem unable to allow someone to politely disagree with your position and will continue heckling comments until either the thread is locked or people just stop responding.

 

Your position is purely your opinion with the only source you've cited being anonymous, other than your disproven claim that the FCC regulates TDPs :lol: . I have testing which has been presented to this community and is freely accessible for anyone to review.

Haswell-DT should overclock well also and hit the same 4.6-5GHz that current 3960X and 3970X can reach without much effort (and remain stable).

 

For flight simulations that can indeed utilize more CPU cores effectively, Haswell-DT may not perform as good as existing 3960X and 3970X 6 cores units.

 

It looks like 2014 is the year when we'll see significant improvements in Intel's CPU performance.

 

Thanks Rob, there are a couple of new bits of info from your report. It's been well understood for a long time re the focus of Haswell (and Broadwell for that matter) being power/GPU related, so that's not news now. Moreover, the times they are a changing. Desktops in the hands of non-business user are fading away in favor of laptops & tablets & and other mobile devices.

 

What was said to imply 'Haswell will overclock well'? This is an area that has been completely hidden so if you have something solid here, that is great news for folks like me who are still on Core 2 Quad or AMD and wish to upgrade.

 

Why will 2014 be the year for significant CPU performance? This sounds a bit iffy when the clear trend is low power, higher GPU, in short CPUs addressing the mobile sector.

 

What simulations can use 6 or 8 cores?

 

If the TIM on the Haswells is good and the chip doesn't run too hot

 

Doubtful here, maybe. With IB-E chips, with their superior TIM, being delayed until after Haswell-DT, what's the basis of thinking Haswell will resolve the TIM issue?

This fall I will (hopefully) be upgrading to Haswell, been looking forward to it for some time to retire my aging Lynnfield build. A new socket, new architecture, better performance, what exactly is there not to like about it other than it's not here yet. I'm still miffed that Intel ditched the 1156 socket without a worthy quad core upgrade; I doubt that 1150 will be a repeat of that.

I'll be interested to see what Haswell really does bring to the table. Hopefully enough people here will buy these new CPUs and test them properly so we can get a good idea. Maybe Word Not Allowed could get ahold of one.

If I haven't gotten a second 3770k (or higher, if they release such a chip) by the time Haswell is out, I'll go Haswell.

What was said to imply 'Haswell will overclock well'?

 

I said nothing to that effect, Broadwell is the CPU that "should" provide more headroom for overclocking. I should clarify the 10% improvement moving from a Sandy Bridge-E 3970X (32nm TDP 150W - 6 core 25W per core) to Haswell-DT and does NOT include any overclocking. Haswell-DT is 22nm (TDP 84W - 4 core 21W per core), Broadwell-D (TDP 65W - 4 core 16W per core) is 14nm. But it's also important to note the L2 and L3 Cache size differences from 3970X (15MB) to Haswell and Broadwell.

 

I should also clarify the 30 fps to 33 fps is for FSX and is just a generic example not a definitive result or what I currently average on my FSX setup.

 

Haswell-DT should overclock a little better than current 22nm Ivy Bridge units.

 

However, I do want to stress that my orignal post was not about overclocking and the "best" methods to overclock and what CPUs will overclocker better than others. I perform a relatively conservative overclocking on my computers with simple water cooling as I need my computers to be stable 24/7 (I use them for more than just my flight sim experience, I use them to also render 3D models and animation which in some cases can run for 48-72 hours straight stressing out all CPU cores).

 

Engineering samples have long since been distributed to Intel partners so the March-June should be a real date as motherboard manufacturers finalize designs and ramp up production (I'd venture to say end of April).

 

Why will 2014 be the year for significant CPU performance?

 

Significant relative to going from a 3970X to Broadwell ... should be a 20-25% improvement. Of course "significant" is subjective and I define that by at least a 20% increase in CPU performance. But I should also stress that overall performance of applications/games is more than just CPU, it's motherboard design, memory, I/O, and graphics (for FSX, it's memory, CPU, I/O ... for XPlane it's memory, CPU, I/O, Graphics) - for example Quad Channel memory and SSD improved fluidity of FSX but did NOT improve frame rate.

Hi

I shall uppgrade to Haswell , get some info from a friend .

I not uppgrade for 5-7% that all i can say, you have results published soon .

A jump from SB-E to Broadwell would provide a large increase in performance, but you're also talking about moving from a 2011 era CPU to a 2014 CPU, and jumping 2 product generations in between. Generally speaking, performance increases cited by the manufacturer compare 2 successive generations and obviously offer less drastic increases. Broadwell will not offer significantly more performance than its predecessor, Haswell, just as Ivy Bridge does not offer significantly more performance than its predecessor, Sandy Bridge.

Hi

I shall uppgrade to Haswell , get some info from a friend .

I not uppgrade for 5-7% that all i can say, you have results published soon .

 

I see you've got an Ivy Bridge CPU. I think that in this case, a Haswell CPU is not worth it at all.

I see you've got an Ivy Bridge CPU. I think that in this case, a Haswell CPU is not worth it at all.

 

You don't know Hasse He's doing it out of sheer curiosity, I imagine.

I said nothing to that effect, Broadwell is the CPU that "should" provide more headroom for overclocking.

 

Ahh, but you clearly did in your opening monologue:

 

'Haswell-DT should overclock well also and hit the same 4.6-5GHz that current 3960X and 3970X can reach without much effort (and remain stable). But Haswell is 4 core (no 6 or 8 versions).'

 

You are entitled to retract your comment if you incorrectly stated your beliefs initially.

 

Noel

 

Haswell-DT should overclock a little better than current 22nm Ivy Bridge units.

 

Really? How do you know this? Is this from Intel engineers or from what source do you base this claim on?

 

Significant relative to going from a 3970X to Broadwell ... should be a 20-25% improvement. Of course "significant" is subjective and I define that by at least a 20% increase in CPU performance.

 

So that's roughly 10% bump per development cycle which is within the realm of recent history. Be aware it's to alleged to be BGA only, i.e., soldered to the main board for what that's worth. When all's said and done, my guess will be that Broadwell will offer about as much over Haswell as Haswell offers over IB, except you won't be replacing a fried CPU as easily ;o)

Noel,

 

No retraction needed, you appear to be mixing Haswell 22nm with Broadwell 14nm. You also have projected significant performance improvement as meaning how well a CPU overclocks. Which is not what I wrote. Both 32nm and 22nm hit similar thermals per core and both easily overclock to 5gHz. Since Haswell is still 22nm with slightly better thermals, it'll hit similar overclocks.

 

Please re-read what I wrote carefully and your own questions. A CPUs performance is much more than its ability to overclock. Moving to Broadwell gains the new design and also provides more significant drop in thermals/energy.

 

My info comes from working with an engineering sample and that's as much as I can say for now.