Forget Intel Broadwell, Skylake On the Way

this doesn't have the FIVR though so it might help issues.

On a different topic i still ahven't seen a good comparison of how the FIVR works. I made a detailed excel sheet of IB scalability and i haven't got my hands on a haswell chip to see if FIVR did anything. I am seriously interested still in how FIVR helps single core load power draw

 

Yes seriously, Charles is right. So many people do not understand that TDP is a specification. It's based on Intels own software testing for worse case scenario barring power viruses. Sure, there might be some variance due to manufacturing not being a perfect process.

While still operating at default settings? If you are not using the default settings out of the box then you are operating out of specification so how can you expect TDP spec to be meaningful anymore. For instance if you were to use any of the 4 overclocking bins on your 3770 or change the default voltages, you would be operating out of spec.

By "this" do you mean Skylake? If so then it's just speculation unless you know for sure that Skylake will not be using IVR for individual domains. i.e. IVR of the cores, IVR for the iGD each with independent supplies from the board for separate current measuring and not having to use one input voltage level for all domains thereby helping with efficiency.

Lol, low TDP scares me The 95W info has been available for several months. Don't forget TDP is a "package" specification and includes all domains including the much bigger iGD's, not just cores. The specs back then for the 95W TDP had core TDC set in the 60's so maybe only up to 50W or 60W for core power at defaults while the iGD TDC may be as high as 90A. Besides, there still time for changes to the arch'.

 

The TDP acts more as a governor. Plenty of chips throttle at their rated TDP. The chips can do much much more then their rated TDP very safely. No one really pays attention to the TDP when designing a cooling system. TDP is meaningless beyond the limit it governors the CPU at. Beyond that the TDP means nothing. Intel is well know for selecting TDP ranges that do not make any real sense when it comes to "power virus". They make TDP settings just because or to limit a CPU so they can have more products like P&G.

go read the wiki page on what CPUs have what TDP because you don't clearly understand Intel's classes for TDP. They also are not done in any major methodological way.

keep up
Intel to Abandon the Internal Voltage Regulator (IVR) with Skylake Microarchitecture

 

You mean manufacturers, surely not?

Details of TDP can be found under the TMSDG's freely available from Intel.

It's probably only using 40W because you are not fully utilizing the processor.

Here's my 3770k set to default so it runs the same as your 3770. Only change is to PL1/PL2.


TDP is 77W so pretty close with 82W. Obviously Intel have their own way of testing to get 77W. The TDP limit is there to keep the processor in thermal spec. If the maximum local ambient temperature spec is 45C and your running at 25C then sure, that gives you an additional headroom of 20C, but the spec is for worse case.

That does not confirm that the IVR is no longer on Skylake. It may be right it may be wrong, I have all ready given a reason why.

 

Man, the wait for Skylake is killing me. I really hope the mainstream/performance lineup is out in time for holiday 2015, and things aren't delayed again.

 

Same. I was expecting a mobile skylake computer to launch in Q1 or Q2 2015, not Q4 2015 or Q1 2016

 

Digitimes Research: Skylake processor delay to weaken Windows 10 notebook demand

 

And so it begins, as we all knew it would. IDK, as long as I can get a new MacBook with Skylake by the holidays (and hopefully not just the Airs), I'll be happy.

 

So, how reliable is this?

From the link in post 207.


And they know today that Skylake will be delayed a year from now? Seems like they simply ran out of real 'news'.

 

Digitimes is like the AP of the computer/electronics supply chain world. I used to reference them quite a bit when I wrote a "News Bits" feature for NotebookReview many years ago. They are usually the first to publish certain info.

 

You could pull that info out of your..errr... behind... and be right... Intel=Delay. I predict at least another delay for Skylake and a release around Q1 2016. You heard it here first.
I'd love to be wrong and have it released sooner..

~Aeny

 

I love that, after all these "forget..." threads, this time around we've actually forgotten about broadwell because it's been delayed so much and skylake is planned to be released so soon afterwards.

 

About how long until broadwell or skylake mobile cpus start showing up for configuration on boutique sites?

I really want to grab that new P650SE or SG, but not if a cooler cpu is right around the corner.

 

The lowest-power (ultra low voltage) Broadwell processors will be out within the next month or two. Then low-voltage Broadwell should be released in 3-4 months from now and standard-voltage Broadwell (including all quad-cores) in 6 months from now. Then, next year around this time, Skylake should be released.

 

Launch schedule of Intel mobile CPUs

 

I'll take the Haswelll Quad Core junk over the broadwell BGA junk anytime.. No harm done if Haswell M processors stay on till Q4 2015...

 

It looks like the general trend is BGA only...

 

Q3-Q4 for Quad cores? Is that what I'm reading? What?

 

Unless you look at the new Clevo P771...

 

Ohh guys, earlier was said, Brodwell U series are available about at this time frame. Now you saying I need to wait another 2-5 months?

 

When has Clevo been associated with the general trend.

 

Never but if you don't want BGA crap, the new P751/771 are the in thing IMO with a PGA Desktop CPU

 

Dell skips Broadwell with some laptops, plans for Skylake | PCWorld

 

OK. I'm starting to get irrationally excited for Skylake.

 

New Wintel hard to drive PC replacement, say Taiwan makers

 

Intel confirms Skylake on time

 

Back to School says to me at least the low-power ultrabook parts. Maybe mainstream. Just give me quad-core mobile Skylake in 2015, Intel, dammit!

 

Intel Confirms Skylake in 2H 2015 – 10nm Still Far Off

 

Thank goodness!! Plenty of time to find one last non-Skylake machine b4 then!! I thought 2.4GHz electromagnetic waves was bad for your health!! 5GHz is stretching into the danger zone!! 60GHz is just plain nuts!! Anybody remember the time when x-ray machines were readily available at shopping malls in the 60's/70's?!?!

Same analogy, only now they're laptops and tablets rather than x-ray machines!!

I'm outta the computer hardware industry after this, no more computer systems that emit anything more than 5GHz!!

 

OMG, we have been under 6*10ⁱ⁴ Hz radiation for so long! We are killing ourselves with such ignorance! We should start going underground right now! I knew I spent so much time mining in Minecraft for a reason!

 

That's what they said when they introduced coin-operated x-ray machines at shopping malls in the 60's/70's --- We're so constantly bombarded by x-rays from outer space, might as well mine some "coins" getting just a little more exposure at the malls... Then, BAM! 20 years later, no one ever thinks of providing x-rays for entertainment ever again d/t health hazards...

Anyways, I will be going into computer hardware hibernation after this year for a very very very long time partly b/c the 60GHz capability will be built onto the system chip... Not a problem if I could discard the wifi card and excise 60GHz from the system, but no they had to incorporate it into the CPU/SoC!!

 

I'm fairly sure the 802 wireless standards have been around 20 years now, and the trend continues toward greater deployment. Using the X-ray argument, don't you think there would be moves to reduce exposure if wireless signals were dangerous?

I remember when there were studies that showed higher levels of certain types of radiation released by cell phones could be harmful, and the manufacturers all very quickly reduced the levels of that type of radiation emitted from their phones. From that as well as the X-ray example, it's clear the industry is responsive when there is a clear and legitimate risk.

Since there has been no reactions or reductions in wifi use, and such use has been going on for quite some time, it's very likely to be safe. Also, unless you live and work in a very sparsely populated area, you're probably getting more "exposure" to wireless frequency radiation from other people's devices than from your own.

 

All the studies done up to now mostly done on 2.4GHz, to a lesser extent 5GHz. This is friggin 60GHz new operating spectrum we're talking about!! It's just nuts to think I would want this in my device!!

 

Given the much higher frequency, the waves will have a proportionally smaller wavelength. This matters because the shorter wavelength signals can't penetrate things like walls (or skulls) - as a result you get high throughput, but line of sight is needed.

This suggests that the waves probably can't get far enough into your body to mess things up.

#Physics

Edit: I did one 5 second google search and found this -study from 2011 (4 years ago - yes WiGig has been in development/discussion for at least that long): http://www.cemmedic.hu/uploads/dokument/Millimeter-wave Scientific honlap.pdf

 

Uhhh I'm not sure if that's entirely true alex, both gamma and X rays have extremely short wavelength but both are very penetrating, requiring very dense materials like lead to stop their propagation.

 

Yes. Also UV radiation doesn't pass through the body, but it still causes skin cancer.

 

Then you have alpha radiation which can be stopped with literally a piece of paper, but if ingested or inhaled it does severe damage to your body.

So yeah, just because something is non-penetrating doesn't mean it's harmless.

 

Touche.

Maybe it's a question of a middle ground/happy medium?

2.4 Ghz ~ 125,000,000 nm wavelength
60 Ghz ~ 5,000,000 nm wavelength
Xray - 0.01-10 nm wavelength ( X-ray - Wikipedia, the free encyclopedia)

(also used Wavelength Frequency Calculator and Microns to Nanometers Conversion Calculator)

 

And 92-95GHz is on the way: UK Based TWEETHER Project Promises 10Gbps mmW Wireless Broadband - ISPreview UK

 

Mark my word, I will never have anything beyond 5GHz beamed from within my laptop, tablet, or cell phone... Even then, where possible, I will limit the 5GHz and go on 2.4GHz instead on a dual band setup, where possible...

 

Will you also be living in isolation in the woods, or do I see a tin foil hat in your future?


Sent from my iPhone using Tapatalk

 

It'll be fine, stop worrying so much about the gigahertz. Bodies are made to be destroyed anyway.

 

To put this in perspective, a 5GHz wave has a wavelength of 60,000,000 nm, while the longest X-ray has a wavelength of 10 nm. They're different by a factor of 6 million. SIX MILLION. So yeah relax.

And in case you didn't know, all these GHz frequencies fall squarely into the "microwave" category. Might want to stop using that popcorn machine if you're this concerned...

 

Intel's Core M 'Skylake' chips coming later this year

 

It smells like no substantial performance increase with Skylake... F-word!

 

http://wccftech.com/intel-skylake-benchmarks-leaked-sisoftware-sandra/
Well, if you believe these "leaked" benchmarks, it looks like Skylake will have about a 20% IPC improvement over Haswell.

 

The whole idea with Broadwell and Skylake is more energy efficiency.. Performance improvement is second fiddle... And CPU performance improvement is 4th fiddle because 3rd fiddle is GPU performance improvement if we see what has been done in Broadwell.. I couldn't care less about iGPU performance other then it being able to handle 120Hz screens or output 4K etc.. I don't need any more performance compared to my HD4000.. I have my 970M and dedicated GPU to do that.. What I want, is more CPU performance ffs..

 

TomJGX,

That is only true of what has been released so far. 20% IPC improvement (if true) will make Skylake a great upgrade on higher TDP platforms which will be able to take advantage of the higher performance while also being much more efficient at idle/near idle too.

Myself, I don't care about GPU's. They're power hungry, they're loud and they heat up the rest of the system for no good reason - Everything Intel is doing with each platform generation lately is geared for me.

They are improving everything they can at once and giving me the platform of choice for many decades now. Lower power is important for all mobile users. And is especially welcome when that comes even with modest core/raw performance improvements each time.

The (new) performance monsters have yet to be released. But there is no doubt they are coming.

 

Intel's track record of mobile platform releases has with few exceptions been very consistent, with each successive platform providing a mild to moderate performance boost with power consumption equal to or less than that of the outgoing platform. So to say that Intel is prioritizing power consumption and disregarding everything else is a little misleading. Indeed they are prioritizing power consumption but they have to balance that priority with other priorities like performance improvements to give the market the right compromise.
Additionally keep in mind that Skylake is more than a processor upgrade. It's going to bring new chipsets, DDR4 memory support, wireless/LAN upgrades and other technologies into the equation. There's a good summary here.

See the PC World link @Jayayess1190 posted on the previous page - it says Skylake is going to be Intel's most significant processor design of the decade. That's significant in itself looking back at what has happened over the last decade in processors alone.

I'm avoiding saying Intel can't do wrong - they can (*cough* Prescott *cough*) - but based on all they've demonstrated so far, I think Skylake will be worth the wait.

 

DDR4 is a necessary progression of memory technology so that we can eventually get higher speeds and density, but for at least the next couple of years the only real benefit will be to people who require 16GB SO-DIMMs for 32 or 64GB mobile configurations. There aren't a whole lot of power savings that can be squeezed out of 1.3v DDR3L, and on the high bandwidth end we already have 2400Mhz DDR3 modules. DDR4's mainstream release will be with 1600Mhz 1.2v modules. Just because Intel wants to combine the release of new processors with new wireless cards doesn't tie them together in any way, otherwise I wouldn't be able to run a 7260AC with my Core 2 Duo.

Power consumption has definitely taken a priority in Intel processors. Low voltage processors used to be rare, but now they are the most common processors in new laptops. It is kind of sad that there are current Haswell and Broadwell flagship "Core i7" processors being sold which are weaker than mobile processors that came out 5-6 years ago. When taking into account that many new processors have their turbo boost features considerably weakened, there are a whole slew of currently sold laptops that will perform most tasks slower than a freaking Core 2 Duo, and people wonder why the market for new computers keeps shrinking.