Intel Skylake / Kaby Lake

[Glo.]

Yes there are three 28W TDP HD6100. Can you find and post any benchmarks between the 28W TDP HD6100 and the 35W TDP HD530 ??




Ehm, where that 80% over Skylake Iris Pro comes from ??? And what Skylake are we talking about ??? GT3e ??

Notebookcheck is full of them. You can find them out easily.

If you would read properly what I wrote: IF the RUMORS are true about Kaby Lake GPU being 80% faster than Skylake(Iris Pro vs Iris Pro). So far rumors about Skylake GPUs are quite true. I genuinely hope that rumors about Kaby Lake will be also true.

 

[Rngwn]

Out of curiosity, those performance estimates are for synthetics (like 3D Mark) or actual games ?? Because Intel iGPUs have way different performance in actual games than on synthetics.

Synthetics (3DMark Firestrike, specifically), sadly. We will need to wait for NBC to actually test the GT3s/GT4s with games. The only gaming benchmark we have by NBC is Bioshock Infinite on SP4, but that should not be used to compared with any benchmark elsewhere, because the definition of "low", "medium", or "high" does not mean the same setting across the reviewers.

IF the RUMORS are true about Kaby Lake GPU being 80% faster than Skylake(Iris Pro vs Iris Pro).

A link to the rumor would be nice As far as I can tell, SKL GT2 and KBL GT2 have the same amount of EUs (24), so I don't expect miracles.

Source indicating that GT2 KBL probably has 24 EUs: http://ranker.sisoftware.net/show_run.php?q=c2ffc9fcdabbdae7deebdcecd8fe8cb181a7c2a79aaa8cffc2f2&l=en

 

[Sweepr]

Some numbers comparing Skylake-U to Carrizo from AnandTech's review.
Despite the fact that Intel closed the gap in graphics performance, most of the time AMD still loses badly when it comes CPU performance:

www.anandtech.com/show/10000/who-controls-user-experience-amd-carrizo-thoroughly-tested-x4

To those who think that AMD has a significant (graphics performance) advantage over Intel when it comes to notebook chips. NotebookCheck just tested the full blown Pro A12-8800B APU (fastest Carrizo, 512 SPs @ 800MHz - FX-8800P equivalent) with dual-channel RAM. Core i5-6200U is one of the most popular Skylake-U models, based on regular HD Graphics 520 (GT2, not fancy GT3e with eDRAM).

In the games where both chips were pitted against each other (19 games total), Core i5-6200U was faster in 10 games while A12-8800B was faster in 9 games at 1366x768 using medium quality settings.

http://forums.anandtech.com/showpost.php?p=37974387&postcount=6229
www.notebookcheck.com/Test-HP-EliteBook-745-G3-Notebook.157955.0.html

 

[Glo.]

A link to the rumor would be nice As far as I can tell, SKL GT2 and KBL GT2 have the same amount of EUs (24), so I don't expect miracles.

The rumor about 80% faster KabyLake GPU is about Iris Pro only.

 

[nvgpu]

http://www.anandtech.com/show/10000/who-controls-user-experience-amd-carrizo-thoroughly-tested-x4/16

Carrizo gets utterly CRUSHED by Skylake in performance, CPU or GPU.

 

[Arachnotronic]

The rumor about 80% faster KabyLake GPU is about Iris Pro only.

I don't think Kaby Lake GPU is supposed to be 80% faster than SKL. I think this figure, if correct, would be a comparison of KBL to BDW.

 

[dark zero]

http://www.anandtech.com/show/10000/who-controls-user-experience-amd-carrizo-thoroughly-tested-x4/16

Carrizo gets utterly CRUSHED by Skylake in performance, CPU or GPU.

Crushed on CPU, but barely winning on GPU... However, seems that Intel is targeting nVIDIA now since the Iris GPU is trading blows against the GT 940M which must be similar to the GTX 750 (no Ti), so... Seems that the gap on GPU on AMD iGPU will be bigger than expecting.

 

[Abwx]

Some numbers comparing Skylake-U to Carrizo from AnandTech's review.
Despite the fact that Intel closed the gap in graphics performance, most of the time AMD still loses badly when it comes CPU performance:

www.anandtech.com/show/10000/who-controls-user-experience-amd-carrizo-thoroughly-tested-x4




http://forums.anandtech.com/showpost.php?p=37974387&postcount=6229
www.notebookcheck.com/Test-HP-EliteBook-745-G3-Notebook.157955.0.html

In all those tests Carrizo is set at 15W in the bios, what about the Intel chips..?.

As i posted it in the other thread the 5200U is roughly at 25W boost in thoses tests, the SKLs are even higher and for a longer time.

I guess that s what it takes to show Intel in good light, unlimited TDP for this latter and reduced as well as controled 15W TDP for AMD, anyway a quite pathetic review that show a complete lack of professionalism.

 

[nvgpu]

http://www.geforce.com/hardware/notebook-gpus/geforce-940m/specifications

GM108 is not a GM107, yet this guy keeps spouting bs like "GT 940M which must be similar to the GTX 750 (no Ti)"

Try again when you actually use your brain to think.

 

[Arachnotronic]

anyway a quite pathetic review that show a complete lack of professionalism.

Wish you'd stop polluting these boards with this nonsense. Anyway, you never answered my question about your l33t engineering training and the big bucks that you are paid to perform engineering work.

 

[dark zero]

http://www.geforce.com/hardware/notebook-gpus/geforce-940m/specifications

GM108 is not a GM107, yet this guy keeps spouting bs like "GT 940M which must be similar to the GTX 750 (no Ti)"

Try again when you actually use your brain to think.

I ommited 2 words.
"GT 940M which must be similar to the GTX 750 (no Ti) on performance"

If I didn't ommit those words, the misunderstanding didn't happen.

 

[Abwx]

Wish you'd stop polluting these boards with this nonsense. Anyway, you never answered my question about your l33t engineering training and the big bucks that you are paid to perform engineering work.

Pollution is from people answering with an ad hominem to technical statements, exactly what you did.

As for electronics i did analog design for a living, and in this domain what matters are the transistors core caracteristics and the related understanding, for the rest i let you speculate on fairy taled TDPs wich ignore this basical law :

Now before getting further in doubting about my credentials, could you comment what you are understanding from the formulae above..?.

As a hint it s the current through the transistor, you know, those Id Max/Id Sat that you once commented as being better for a firm than for another, so let s see if you knew what you were talking about at the time.....

 

[nvgpu]

You're still wrong, GM108 GT 940M only has 384 CUDA Cores and 64bit memory bus DDR3, GM107 GTX 750 has 512 CUDA cores and 128bit memory bus GDDR5, the performance is NOT even close or comparable. If you're that clueless, don't post.

 

[Sweepr]

I guess that s what it takes to show Intel in good light, unlimited TDP for this latter and reduced as well as controled 15W TDP for AMD, anyway a quite pathetic review that show a complete lack of professionalism.

You can complain and twist the facts/numbers for Intel chips as much as you want, there's plenty of results out there. 35-42W Carrizo with dual-channel is slower than Skylake-U, while 15W simply gets demolhised in CPU benchmarks (especially ST). It's not only a matter of OEM implementation when it still lags behind inside the reference platform with unlimited TDP.

- Cinebench R15 CPU MT 64-bit

Core i5-6300U (Surface Pro 4): 309
Core i7-6650U (Surface Pro 4): 342
Core i7-6567U (Vaio Z Flip): 373
35W A10-8800P (Lenovo Y700): 282


- Cinebench 11.5 CPU ST 64-bit

Core i5-6300U (Surface Pro 4): 1.4
Core i7-6650U (Surface Pro 4): 1.51
35W A10-8800P (Lenovo Y700): 0.96


- Cinebench 11.5 CPU MT 64-bit

Core i5-6300U (Surface Pro 4): 3.4
Core i7-6650U (Surface Pro 4): 3.8
35W A10-8800P (Lenovo Y700): 3.36
Carrizo (Reference Platform):

According to Shintai Core i3-6100U uses ~8W for a 2.74 score, let's take a look at 15W+ Carrizo:

See you next year AMD.

Wish you'd stop polluting these boards with this nonsense. Anyway, you never answered my question about your l33t engineering training and the big bucks that you are paid to perform engineering work.

I'm afraid you won't get an answer.

 

[Abwx]

You can complain and twist the facts/numbers for Intel chips as much as you want, there's plenty of results out there. 35-42W Carrizo with dual-channel is slower than Skylake-U, while 15W simply gets demolhised in CPU benchmarks (especially ST). It's not only a matter of OEM implementation when it still lags behind inside the reference platform with unlimited TDP.

- Cinebench R15 CPU MT 64-bit

Core i5-6300U (Surface Pro 4): 309
Core i7-6650U (Surface Pro 4): 342
Core i7-6567U (Vaio Z Flip): 373
Carrizo (Lenovo Y700 - 35W A10-8800P): 282


- Cinebench 11.5 CPU ST 64-bit

Core i5-6300U (Surface Pro 4): 1.4
Core i7-6650U (Surface Pro 4): 1.51
Carrizo (Lenovo Y700 - 35W A10-8800P): 0.96


- Cinebench 11.5 CPU MT 64-bit

Core i5-6300U (Surface Pro 4): 3.4
Core i7-6650U (Surface Pro 4): 3.8
Carrizo (Lenovo Y700 - 35W A10-8800P): 3.36
Carrizo (Reference Platform):

See you next year AMD.

What is amazing with you is that on one side AMD power measurements are provided in a controled environment like SA s Thevenin graph that you are displaying, but on the other side Intel s numbers are the ones published at Arkintel and you never provide any real world measurements..

So much for your credibility based on such slight of hand, provide us the same curves for Intel, there s none, and not by chance...

I'm afraid you won't get an answer.

You will be afraid that it s actualy my answer that will find no further question, and neither any answer about my subsequent question....

 

[Arachnotronic]

Pollution is from people answering with an ad hominem to technical statements, exactly what you did.

As for electronics i did analog design for a living, and in this domain what matters are the transistors core caracteristics and the related understanding, for the rest i let you speculate on fairy taled TDPs wich ignore this basical law :

Now before getting further in doubting about my credentials, could you comment what you are understanding from the formulae above..?.

As a hint it s the current through the transistor, you know, those Id Max/Id Sat that you once commented as being better for a firm than for another, so let s see if you knew what you were talking about at the time.....

Attempt at a snow job, as usual. You can stick a big picture of an equation all you'd like, but I don't see any actual numbers to plug into said equation provided by you, making all of your pontificating all for nothing.

 

[Abwx]

Attempt at a snow job, as usual. You can stick a big picture of an equation all you'd like, but I don't see any actual numbers to plug into said equation provided by you, making all of your pontificating all for nothing.

Intel provide no measurements, neither do the reviewers if we except a few reviews with poor explanations about the test condition, in the waiting my estimations are much more credible than the ones extracted from marketing slides, the only numbers you re using in your own "estimations"..

But you have a i7, make a few tests, or where they already done and are you waiting to post them that they are more complete..?..

 

[dahorns]

Intel provide no measurements, neither do the reviewers if we except a few reviews with poor explanations about the test condition, in the waiting my estimations are much more credible than the ones extracted from marketing slides, the only numbers you re using in your own "estimations"..

But you have a i7, make a few tests, or where they already done and are you waiting to post them that they are more complete..?..

Would it really matter if he ran tests? Wouldn't you just come up with a reason why he was lying and/or screwed up the tests?

 

[ShintaiDK]

But you have a i7, make a few tests, or where they already done and are you waiting to post them that they are more complete..?..

Didn't I show you some i7 tests? That completely destroyed your made up fantasy numbers.

 

[Arachnotronic]

Didn't I show you some i7 tests? That completely destroyed your made up fantasy numbers.

The answer lies in the "transconductance." Duh.

 

[ShintaiDK]

The answer lies in the "transconductance." Duh.

Haha

 

[FlanK3r]

Guys, who are Minecraft gamer? What about Mincecraft at 5.7+ GHz ? Tested with 2x GTX980Ti SLI

 

[IntelUser2000]

IF the RUMORS are true about Kaby Lake GPU being 80% faster than Skylake(Iris Pro vs Iris Pro). So far rumors about Skylake GPUs are quite true. I genuinely hope that rumors about Kaby Lake will be also true.

I bet the rumors are just misinterpreting what Intel slides are claiming.

Iris Pro 6200 claim was that it was 80% faster than Haswell. Haswell what? Well 6200 is only 20% faster than 5200 so its against GT2 Haswell, HD 4600 graphics.

Likely Kabylake GT3e is 80% faster than Skylake GT2, or Kabylake GT4e is 80% faster than Skylake GT3e. Either way, not that impressive. That means a worst case 10% and a best case 20%.

However, seems that Intel is targeting nVIDIA now since the Iris GPU is trading blows against the GT 940M which must be similar to the GTX 750 (no Ti),

Oh, come on! Iris Pro 6200 on a 65W Desktop part needed to be 50% faster to catch up to a GTX 750. Meaning Skylake GT4e = GTX 750(not a Ti). You are talking about a previous generation 15/28W GT3e.

It isn't a best case.

You can't be sure. I have looked at all the slides before I posted that. Have you? Up to 50% on a marketing slide means best case.

http://www.pcgameshardware.de/Skyla...erformance-Portfolio-1158623/galerie/2371544/
http://www.windowscentral.com/sites.../image/2015/07/fanless-tech.png?itok=4IM9zsZ0

-50% more slices = 40% improvement(no scaling of front end)
-Lower clocks = less performance
-Better architecture = 20-30% of Skylake over Broadwell
-Impact of eDRAM changes unknown(bandwidth/arrangement)

http://forums.anandtech.com/showpost.php?p=38008140&postcount=6401

1.15GHz Iris Pro 6200 to 1.05GHz Iris Pro 580.

Not to mention not a single manufacturer is using Iris Pro 6200 on a mobile.

Again, the chips are available, so it's more like a 6+ months lead. Some design wins (HP comes to mind) were announced last year, technically they could be launched now.

No its not. You can't buy them. And Polaris is mid-2016, which makes it CES. Too bad for Intel, it used to be devices were available 1-2 months of launch, nowadays with some products you can't even buy them at all.

And "H1" is quite ambiguous. They say H1 rather than Q1 because they aren't sure they can't meet Q1 and has to be after or they just want to confuse speculators. Iris Pro 6200 mobile parts came out June last year. It'll be Late April at the earliest for Iris Pro 580. And vast majority of manufacturers(like everyone but Apple) will continue to use discrete graphics and likely wait 2 months the much better Polaris and ignore Iris Pro 580. Unless Intel does something radical. Like not be greedy.

 

[Abwx]

The answer lies in the "transconductance." Duh.

Lol, you dont even know what that means, i guess that making fun with enginering terms is just aknowledgment of bad faith.

Anyway keep on spreading your marketing TDPs on a site that is supposed to be about actual enginering facts, surely that such facts are incompatible with some quasi religious beliefs when it comes to some brand...

 

[Arachnotronic]

Lol, you dont even know what that means, i guess that making fun with enginering terms is just aknowledgment of bad faith.

Anyway keep on spreading your marketing TDPs on a site that is supposed to be about actual enginering facts, surely that such facts are incompatible with some quasi religious beliefs when it comes to some brand...

Hey man, if you think you know more than what the review sites are telling us, I strongly encourage you to use this knowledge in order to make better purchasing decisions. If you think AMD makes the best product, then it's not my place or anybody else's place really to tell you what to buy.

Buy AMD products. Enjoy AMD products. It's your money and if buying AMD gives you the satisfaction that you have the best possible product on the market, then AMD powered products are the way to go for you.

And if once you buy those products, you find that they do everything that you wanted and more, then that's great, money well spent! I have ZERO objection to that (and even if I did, who cares, it's your buying decision and yours alone)!

Here's what I do have an objection to, though. By pretty much all objective third party tests, and given the design wins Intel has been able to score, Intel right now makes the better product in many cases. Their media support is unmatched in PCs (mobile SoCs will actually pull ahead though soon), their CPU performance is first rate and power consumption/battery life of Intel powered products is very good.

There's a reason that OEMs choose Intel and it's not because they're being "bribed" -- it's because Intel makes the better product and provides a level of platform/device level support that AMD simply doesn't have the resources to match.

But you come in here, accuse Intel of lying/misleading their customers, and you try to downplay the very real value that Intel delivers to its customers with its best-in-class products.

You then throw around a lot of technical jargon that supposedly "proves" that Intel is all marketing fluff and no technical substance, but IMO your arguments don't actually make any sense because you basically hint at stuff and hide a lot behind "oh here's a fancy formula, if you were an EE you'd get it."

I leave this quote for you to ponder:

If you can't explain it to a six year old, you don't understand it yourself.