Intel Skylake / Kaby Lake

Page 256 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Abwx

Lifer
Apr 2, 2011
10,847
3,297
136
Yeah, ok. Whole world is being duped by Intel marketeers...what a joke.

Buy what you like but don't try to lead people into making bad purchases just because you don't like Intel.

Of course you pay no attention to thoses "15W" Intel mobile CPUs that score anywhere from 2.2 to 3.8 in cinebench, nevermind that the power difference is automaticaly a 2.7 ratio given the frequency ratio, we are told that it s all "15W" TDPs...

Say what you want, twice a score in CB require close to twice the power if cores are doubled, but 4X the power if relying on frequency with a same core count.

As said ad nauseam if a 2C/4T can score 3.8 at 15W then it wouldnt require more than 27-28W for a 4C/8T to score 7.6, yet all Intel DT parts that score 7.6 use way more than this power for such numbers.

For you thoses numbers mean perhaps nothing because you have apparently no formal enginering training, but for me when i hear some claims by here it s like hearing someone saying that one and one makes four, so in a way you are right since it s all what is marketing about, inflating the numbers, including of course Intel s...
 
Mar 10, 2006
11,715
2,012
126
For you thoses numbers mean perhaps nothing because you have apparently no formal enginering training, but for me when i hear some claims by here it s like hearing someone saying that one and one makes four, so in a way you are right since it s all what is marketing about, inflating the numbers, including of course Intel s...

All right, what formal engineering training do you have? What company currently pays you for your engineering expertise?
 

ShintaiDK

Lifer
Apr 22, 2012
20,378
145
106
6700K uses ~58W for a 9.66 CB score. 6100U uses ~8W for a 2.74 score.

And those are IA Cores+GT+Uncore.

Not that the compare can be used for anything.
 
Last edited:

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
s said ad nauseam if a 2C/4T can score 3.8 at 15W then it wouldnt require more than 27-28W for a 4C/8T to score 7.6, yet all Intel DT parts that score 7.6 use way more than this power for such numbers.

There's a review by Anandtech which shows power information of a 2nd Gen Core i7 quad core running games. While a 15W essentially is at 15W all the time, the 45W part isn't. That's the *key*.

45W for a quad core is for three scenarios:
-CPU-only, but extremely demanding
-GPU-only
-CPU + GPU

It's quite likely Cinebench isn't the most demanding program out there. You talk about "Law of Physics", fortunately plenty of wiggle room exists within the "Law" to make widely different products even in the current era of extremely diminishing gains. Unfortunately for AMD there's the problem of "execution". While its true Intel is doing shady things and some things are extremely inflated(like Core M, still), using AMD as an example is really not the best thing to do.

Also, there's the aspect of binning. Have you forgotten?

*Sweepr*

Regarding Core M Skylake, they still haven't found a way to make the CPU beyond the cheapest Core m3 version perform better. Notebookcheck's results show m5 and m7 are barely faster(~10% maybe). That's why extremely limited amount of manufacturers use Core M, and even fewer for m5 and m7, because they offer no benefits.
 
Last edited:

NTMBK

Lifer
Nov 14, 2011
10,208
4,940
136
Regarding Core M Skylake, they still haven't found a way to make the CPU beyond the cheapest Core m3 version perform better. Notebookcheck's results show m5 and m7 are barely faster(~10% maybe). That's why extremely limited amount of manufacturers use Core M, and even fewer for m5 and m7, because they offer no benefits.

When you're clamped to a hard 4.5W power wall, there's only so much you can do. As soon as any sustained load kicks in, all of those parts are going to throttle down to the same speeds. Your brief turbo performance spike will be a little faster, but you're also going to hit thermal limits even quicker.

(For reference, I have an i5-4300Y tablet, which has a 6W power limit set in the BIOS. It throttles like a MF as soon as you load the GPU and CPU simultaneously.)
 

ShintaiDK

Lifer
Apr 22, 2012
20,378
145
106
Just got the new BIOS for my MSI Z170I with the new microcode.

And like the 6100U NUC, the scores is the same with the 6700K in loads. No performance loss. The NUC actually got slightly faster (2-3%), but may be for different reasons.
 
Last edited:

dark zero

Platinum Member
Jun 2, 2015
2,655
138
106
The world is busy buying Intel based laptops and not the AMD laptops that some allege are super and amazing. Gotta be some reason...
The big problem is that Carrizo is so awful that doesn't need to be an expert to detect that.. On the other side... People are buying U chips based laptops instead of H ones...
 

witeken

Diamond Member
Dec 25, 2013
3,899
193
106
Intel ISSCC: http://www.eetimes.com/document.asp?doc_id=1328835

What does this mean:

The general manager of Intel’s technology and manufacturing group declined to share his thoughts about which of a “rich variety” of post-CMOS technologies chip makers will use or when. New techniques span tunneling FETs, ferroelectric FETs, spintronics, new III-V materials and more.

Holt did assert the new techniques won’t be in Intel’s 10nm process in which Intel is now prototyping its next-generation processors. In general, engineers will stretch CMOS as far as possible. Longer term, chips will be hybrids of different techniques blended with traditional CMOS.

Is 10nm just a shrink of 14nm? Or does it have III-V, since technically those materials aren't post-CMOS.
 

dark zero

Platinum Member
Jun 2, 2015
2,655
138
106
Intel ISSCC: http://www.eetimes.com/document.asp?doc_id=1328835

What does this mean:



Is 10nm just a shrink of 14nm? Or does it have III-V, since technically those materials aren't post-CMOS.
Hope that is not a mere shrink, if that so, is very bad news.....

Ps: It happened...

http://wccftech.com/intel-forcing-ban-nonk-oc-feature-skylake-motherboards-bios-rolling/

Intel will lock their non K OC. On all the boards, unless someone screws Intel up and not take that lock, but is unlikely since Intel would block the sales of that boards.

Also that means that the next stepping of processors will come with OC lock in order to cancel the boards who still have that feature available.

Seems that the non K OC heavily damaged their sales. The i5 6400 on OC was better than the 6600K on most aspects, even power consumption.

Too bad that Intel went on Jerk mode since that generation was starting to shine. Also that was the 1st generation about time to see a OC able Celeron, making it competing to even the Core i5 on ST and Core i3 on MT.


With that Intel assassinates one of the key features of that generation and kills the hype generated on that.
 
Last edited:
Mar 10, 2006
11,715
2,012
126
Intel ISSCC: http://www.eetimes.com/document.asp?doc_id=1328835

What does this mean:



Is 10nm just a shrink of 14nm? Or does it have III-V, since technically those materials aren't post-CMOS.

Shrink of 14nm. If I had to guess, Intel will make the fins taller & thinner to deliver more performance and you will see the air-gaps at more metal layers in a bid to lower parasitic capacitance. Also expect fins to be squeezed even closer together.
 

witeken

Diamond Member
Dec 25, 2013
3,899
193
106
Shrink of 14nm. If I had to guess, Intel will make the fins taller & thinner to deliver more performance and you will see the air-gaps at more metal layers in a bid to lower parasitic capacitance. Also expect fins to be squeezed even closer together.

How about SiGe? I mean, IBM has SiGe at 7nm.
 

Rngwn

Member
Dec 17, 2015
143
24
36
Another i7 6567U benchmark on Vaio Z Clamshell (クラムシェルモデル) and Flip (フリップモデル). The entire stuff is in Japanese, if someone wouldn't mind translating it for me.

http://pc.watch.impress.co.jp/docs/column/hothot/20160127_740827.html

Here are what I think is interesting:

Clamshell:

3DMark13 Firestrike: 1523
3DMark13 Firestrike Extreme: 682

Flip:

3DMark13 Firestrike: 1503
3DMark13 Firestrike Extreme: 653

Compared with S11(w/ i7 6500U&HD520):

3DMark13 Firestrike: 905
3DMark13 Firestrike Extreme: 416

NBC's Zenbook UX303 (w/ i7 6500U&940M):
http://www.notebookcheck.net/Asus-Zenbook-UX303UB-DH74T-Notebook-Review.156912.0.html

3DMark13 Firestrike: 1347
3DMark13 Firestrike Extreme: 680
Edit: This one have slight GPU throttling, the actual 940M performance result should be a bit higher.


And with NBC's SP4 i7 (Iris 540):

3DMark13 Firestrike: 1395
3DMark13 Firestrike Extreme: 598

This would put the Iris 550 within the same ballpark as 940M (at least, for synthetics). There is also benchmark on Dragon Quest X in the site as well, but I don't think I could find some good comparison with dGPU.
 
Last edited:

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
When you're clamped to a hard 4.5W power wall, there's only so much you can do.

Yes, I know. It's in between a rock and a hard place, that is, Atom and 15W Core i's. It's still a mediocre product.
 
Last edited:

ShintaiDK

Lifer
Apr 22, 2012
20,378
145
106
Another i7 6567U benchmark on Vaio Z Clamshell (クラムシェルモデル) and Flip (フリップモデル). The entire stuff is in Japanese, if someone wouldn't mind translating it for me.

http://pc.watch.impress.co.jp/docs/column/hothot/20160127_740827.html

Here are what I think is interesting:

Clamshell:

3DMark13 Firestrike: 1523
3DMark13 Firestrike Extreme: 682

Flip:

3DMark13 Firestrike: 1503
3DMark13 Firestrike Extreme: 653

Compared with S11(w/ i7 6500U&HD520):

3DMark13 Firestrike: 905
3DMark13 Firestrike Extreme: 416

NBC's Zenbook UX303 (w/ i7 6500U&940M):
http://www.notebookcheck.net/Asus-Zenbook-UX303UB-DH74T-Notebook-Review.156912.0.html

3DMark13 Firestrike: 1347
3DMark13 Firestrike Extreme: 680
Edit: This one have slight GPU throttling, the actual 940M performance result should be a bit higher.


And with NBC's SP4 i7 (Iris 540):

3DMark13 Firestrike: 1395
3DMark13 Firestrike Extreme: 598

This would put the Iris 550 within the same ballpark as 940M (at least, for synthetics). There is also benchmark on Dragon Quest X in the site as well, but I don't think I could find some good comparison with dGPU.

+1 :thumbsup:

The lower end dGPUs are going out of business fast. Good thing that higher end gaming revenue is up significantly (GTX970 and up).
 

MarkizSchnitzel

Senior member
Nov 10, 2013
399
28
91
+1 :thumbsup:

The lower end dGPUs are going out of business fast. Good thing that higher end gaming revenue is up significantly (GTX970 and up).

Well surely not until these Iris parts can be had in devices ~900€?
i5+940M can be had for such a price, and lower..
 

Rngwn

Member
Dec 17, 2015
143
24
36
Let's hope Intel would being more generous and start giving eDRAM on GT2s in the next few genrations. Not that it gonna happen very soon, but interesting to see what Intel has to answer if AMD would begin using the HBM with their iGPUs.
 
Aug 11, 2008
10,451
642
126
The price is already down there now. There is only a small difference between EDRAM and non EDRAM parts.

Perhaps, but what counts is the price to the end user. Iris pro is still pretty much only available in high end (read expensive) devices, even then, usually as an optional upgrade. Like I have said before, if intel really wants to drive adoption of this they should make it standard on all quad mobiles and a lot more ultrabook chips as well.
 

Burpo

Diamond Member
Sep 10, 2013
4,223
473
126
I already told you guys this :) 10nm is basically a straight shrink of 14nm, 7nm introduces new materials.

Confirmed?
"“I can’t tell you which of these [post-CMOS] technologies will be first or best but when we see this richness [of possibilities]…that provides a wealth of opportunities over the next few years to make the tremendous progress needed in how to architect our parts,”

" Holt noted Intel’s 10nm process will support five voltage threshold levels. The variety of optimization points in a given node will likely increase as post-CMOS technologies are added to the mix."

Holt-key2-x-500.png


http://www.eetimes.com/document.asp?doc_id=1328835
 
Last edited:

jhu

Lifer
Oct 10, 1999
11,918
9
81
Another i7 6567U benchmark on Vaio Z Clamshell (クラムシェルモデル) and Flip (フリップモデル). The entire stuff is in Japanese, if someone wouldn't mind translating it for me.

http://pc.watch.impress.co.jp/docs/column/hothot/20160127_740827.html

Here are what I think is interesting:

Clamshell:

3DMark13 Firestrike: 1523
3DMark13 Firestrike Extreme: 682

Flip:

3DMark13 Firestrike: 1503
3DMark13 Firestrike Extreme: 653

Compared with S11(w/ i7 6500U&HD520):

3DMark13 Firestrike: 905
3DMark13 Firestrike Extreme: 416

NBC's Zenbook UX303 (w/ i7 6500U&940M):
http://www.notebookcheck.net/Asus-Zenbook-UX303UB-DH74T-Notebook-Review.156912.0.html

3DMark13 Firestrike: 1347
3DMark13 Firestrike Extreme: 680
Edit: This one have slight GPU throttling, the actual 940M performance result should be a bit higher.


And with NBC's SP4 i7 (Iris 540):

3DMark13 Firestrike: 1395
3DMark13 Firestrike Extreme: 598

This would put the Iris 550 within the same ballpark as 940M (at least, for synthetics). There is also benchmark on Dragon Quest X in the site as well, but I don't think I could find some good comparison with dGPU.

I'll give it a shot since I'm "learning Japanese" (ie watch a lot of anime).