AMD Carrizo Pre-release thread

[Arachnotronic]

So you re using obviously false claims as an argument for your own agenda..?..

Are you serious when talking of metrics..??.

This term is generaly used with people who know what is precision...

Since you took Phynaz claims at face value, can you point me where are the difference in configuration he allegedly see in this slide ..??.

Could you please tell me which false claims I am using, and of what agenda you speak?

I never claimed a difference in configuration. I merely claimed that Skylake-U should be more competitive than Carrizo where it counts.

 

[Phynaz]

You couldnt help but threadcrap the thread out of desperation.

There s no differences, systems are the same and you would be hard pressed to point what are the difference...

2x4GB 1600MHz DDR3, 256GB SSD, Windows 10, 50Wh Battery

That s for all systems, and look like the Intel brigade did follow suit without even reading the slide, apparently they prefer to rely to someone who see imaginary differences and still are playing the experts, lol....

Where does it say the Intel system has an SSD for the idle runtime? What were the display brightnesses set to? What was the power saving configuration of the systems? Why were different operating systems used in different measurements, do they expect customers to do the same? It doesn't even say the Intel system was a commercially available system.

You would have to be a fool to believe that any company would not set conditions to give themselves an advantage when comparing their products to a competitors.

How about instead of cherry picking slides you link to the entire deck? Also, has AMD released their test files so their results can be independently verified? Please link to those also.

 

[Abwx]

Could you please tell me which false claims I am using, and of what agenda you speak?

I never claimed a difference in configuration. I merely claimed that Skylake-U should be more competitive than Carrizo where it counts.

Ok, sorry if i misinterpreted you, not that the poster that followed this troll post didnt aknowledge that he did agree with the claim, must admit that you were more cautious...

As for the competition, why should SKL be more competitive than Carrizo..?.

And what are thoses metrics that count, you mean the ones that suit Intel whatever they are..??..curious about this...

 

[Abwx]

Where does it say the Intel system has an SSD?

Where does it say that it doesnt use one..?..

You think that they used anything else than a SSD...?.

That they did set brightness at different values to grab a few mW..?..lol...

If Intel was to present such a slide i wouldnt even bother insisting with such arguments, thruth is that you are displaying the usual staunch Intel supporter behaviour, that is, that AMD has no right to best Intel, numbers be damned..

That s about what can be summarized from most of the negative posts in this thread, to your relative defense, and it doesnt happen often, AMD is lacking marketing wise if we are to interpret those slides in all objectivity.

Edit : Dont bother for AMD s numbers, they know well before us how Intel CPU perform and how to measure everything.

They are delivered plateforms during the validation phases, you know, they must make sure that their GPUs will work accordingly with each new Intel CPU...

 

[Arachnotronic]

As for the competition, why should SKL be more competitive than Carrizo..?.

More efficient CPU architecture with a more robust DVFS system, stronger media engine (Skylake is said to nullify Carrizo's HEVC decode advantage, while at the same time bringing on VP9 hardware decode* support which is useful for the popular YouTube video watching use case), more efficient manufacturing process (14nm for the main SoC, 22nm for the PCH v.s. 28nm SoC).

*Source: http://cdn.eteknix.com/wp-content/uploads/2015/04/Skylake-S-Platform.jpg

 

[Burpo]

"Carrizo APUs are the first mainstream processors that support HEVC. Intel’s Broadwell CPUs only support a hybrid model where CPU and GPU cores work together, so you might need a powerful Core i7 processor for smooth playback depending on the bitrate.

Carrizo is the first processor in the world that supports version 1.0 of the HSA standard, which allows a better cooperation of the processor and graphics cores. There is currently no popular software that can utilize it, but certain features were and will be implemented by Intel, so it is a pretty future-proof feature.

AMD wants to be successful with Carrizo in the 15-Watt range in particular, and according to the released data and benchmarks, this seems to be a success.

..new features like the UVD 6 Video Engine including the future-proof HEVC decoder, HSA 1.0 support and the TrustZone security processor. Without a dedicated graphics card, the Carrizo APUs seem to be an interesting alternative to Intel’s Core i3 – i5 in the price range between $400-$700 thanks to the powerful graphics, compute and multimedia performance. Our in-depth reviews will show how the APUs perform in available notebooks. AMD currently expects the launch of new devices from Acer, Asus, Dell, HP, Lenovo and Toshiba between July and August 2015."

Thanks for the notebookcheck link.. Pretty good read..

I hope it's all true and AMD gains some traction in the market at 15 watts. Seems to be what everyone wants..

Supposedly tho, that's where Skylake will shine.. I guess we'll know soon

 

[teejee]

Where does it say that it doesnt use one..?..

You think that they used anything else than a SSD...?.

That they did set brightness at different values to grab a few mW..?..lol...

If Intel was to present such a slide i wouldnt even bother insisting with such arguments, thruth is that you are displaying the usual staunch Intel supporter behaviour, that is, that AMD has no right to best Intel, numbers be damned..

That s about what can be summarized from most of the negative posts in this thread, to your relative defense, and it doesnt happen often, AMD is lacking marketing wise if we are to interpret those slides in all objectivity.

Edit : Dont bother for AMD s numbers, they know well before us how Intel CPU perform and how to measure everything.

They are delivered plateforms during the validation phases, you know, they must make sure that their GPUs will work accordingly with each new Intel CPU...

Saving a few mW by changing brightness? A few watt is more correct.

 

[Abwx]

More efficient CPU architecture

You realize that it s a blank statement backed with no argument..?.

Actualy it s less efficient, the same way HW is less effiient than Beema..

with a more robust DVFS system,

How do you know that it s more robust..?.

Beware, the term robust has an enginering meaning and is not used randomly by people working in the field, try to demonstrate that a system has robust stability, it require much more than a handfull of words...

So i m waiting for a robust demonstration...

stronger media engine (Skylake is said to nullify Carrizo's HEVC decode advantage, while at the same time bringing on VP9 hardware decode* support which is useful for the popular YouTube video watching use case)

To nullify is not to exceed, but i m used to your hype..

, more efficient manufacturing process (14nm for the main SoC, 22nm for the PCH v.s. 28nm SoC).

It s not more efficient at the frequencies of interest, otherwise Carrizo idle power wouldnt be lower than BDW, i suggest that you let those kind of estimations to people who can grasp what the process are worth.

So far there s not a single knowledgeable people that contradicted me about my estimations of Intel 14nm process performances relative to their 22nm, and for a reason, it s not even the level of the first year of university, why bother stating the obvious..??.

*Source: http://cdn.eteknix.com/wp-content/uploads/2015/04/Skylake-S-Platform.jpg

You know that Eteknix is among the few sites that were denied Fury GPU because of obvious viral marketing..?.

Not that it s a problem, quite the contrary since it litteraly hammer the final nail in the coffin of your unsubstancied claims, set apart eventualy for HEVC decoding but with quite an inflation...

 

[ShintaiDK]

People believe in company slides?

Just look at the Fiji launch for the latest bunch of lies. Else we can start go back to all the previous times with the extravaganza promises and no delivery.

It doesnt matter what company makes the slides. Its just not to be trusted as we see over and over again.

 

[Abwx]

Saving a few mW by changing brightness? A few watt is more correct.

A Few watts is the max comsumption of a laptop medium sized screen.

And why changing brightness, you know what it means to be an engineer..?..

It means that this kind of debate not only doesnt exist but is simply unthinkable.

People believe in company slides?

.

I guess that it depend of the person, and of the company...

Personaly i always try to check, and generaly the more it s difficult to check the less the probabiliy of the info to be accurate..

I had to check a lot of reviews to sort out Core M intrinsic perf/Watt, comparitively it was a cakewalk with Carrizo and i had a single CPUZ shot, so as you can read the proof is in the pudding.

 

[ShintaiDK]

A Few watts is the max comsumption of a laptop medium sized screen.

By a few watts you mean 5-10W? With 200nits you may get away with 3-4W.

 

[Abwx]

By a few watts you mean 5-10W? With 200nits you may get away with 3-4W.

The discussion is moot since it has no meaning, it started from someone who implied that AMD engineers could be dishonnest, of course that any stockholder could sue them if the numbers are fake didnt seem to pop up in some brains, yet that s the risk when officialy stating those numbers..

Who wants to make millions $..??..

Edit :

http://forums.anandtech.com/showpost.php?p=37577979&postcount=1735

 

[CHADBOGA]

Indeed since this was achieved with 1.5 node handicap...

That said to not waste this post for trivialities i think that a little more info wouldnt be off topic, or is it..??.

There's no doubt about it, AMD has the most powerful Powerpoint slides the world has ever seen.

 

[maarten12100]

More efficient CPU architecture with a more robust DVFS system, stronger media engine (Skylake is said to nullify Carrizo's HEVC decode advantage, while at the same time bringing on VP9 hardware decode* support which is useful for the popular YouTube video watching use case), more efficient manufacturing process (14nm for the main SoC, 22nm for the PCH v.s. 28nm SoC).

*Source: http://cdn.eteknix.com/wp-content/uploads/2015/04/Skylake-S-Platform.jpg

More efficient in terms of resource usage but no longer in power efficiency it allegedly. AMD allegedly made a chip more efficient aka higher performance/watt than Intel on a node 2 shrinks behind Intel. On a failed architecture. With next to no R&D budget.

Seems like Intel has been slacking to me and everyone still drinking the cool aid. Intel might actually have genuine competition with a new good architecture along with a good node. They can start opening up the ready to use designs they obviously have been holding back from the consumer for years.

By a few watts you mean 5-10W? With 200nits you may get away with 3-4W.

So how come notebooks idle at 5W with the screen on at high brightness while the screen takes 5-10W according to you. Clearly not the case any datasheet will tell you that. Maybe with some 400 ?nits? ('murica candela?) at 17".

Reviews will be here shortly anyways. Let's see what Carrizo can do even configured in a crapbook.

 

[gorion]

First benchmarks from notebookcheck on Carrizo-L.
I'd say disappointing, unless they are at significantly lower power consumption.

http://www.notebookcheck.net/First-benchmarks-for-AMD-A8-7410-Carrizo-L-APU-now-online.146802.0.html

 

[coercitiv]

There's no doubt about it, AMD has the most powerful Powerpoint slides the world has ever seen.

This type of replies is exactly what comes to mind when others poop in Intel related threads. In the end all that remains is a strange odor all over the forum. What strikes me the most is many of those who enjoy taunting like this are actually rather educated, with solid technical background, and fully understand the forum rules and more importantly the clear benefits of abiding by those rules.

I get that part of the motivation behind such replies is the person debating with, and certainly Abwx has a very... abrasive posting style, but you should always remember for every person participating in the conversation there are many more listening, most of which deserve better than thread crapping.

 

[maarten12100]

First benchmarks from notebookcheck on Carrizo-L.
I'd say disappointing, unless they are at significantly lower power consumption.

http://www.notebookcheck.net/First-benchmarks-for-AMD-A8-7410-Carrizo-L-APU-now-online.146802.0.html

Carrizo-L is Mullins HP send me a Carrizo-L notebook it was literally Mullins. HP is incompetent.

 

[ShintaiDK]

So how come notebooks idle at 5W with the screen on at high brightness while the screen takes 5-10W according to you. Clearly not the case any datasheet will tell you that. Maybe with some 400 ?nits? ('murica candela?) at 17".

Reviews will be here shortly anyways. Let's see what Carrizo can do even configured in a crapbook.

At idle there is almost only the screen left.

While, old. Example of the Ipad2 vs Ipad3. Resolution got a big impact as well.

 

[ShintaiDK]

The discussion is moot since it has no meaning, it started from someone who implied that AMD engineers could be dishonnest, of course that any stockholder could sue them if the numbers are fake didnt seem to pop up in some brains, yet that s the risk when officialy stating those numbers..

Who wants to make millions $..??..

We dont exactly lack examples, so dont try and act like it doesnt happen.

Remember Phenom beating Core 2 by 40-50%? Bulldozer? Fiji? Just to mention a few.

And you can do the same with intel, nVidia, Qualcomm, Samsung etc.

Who sued Samsung over benchmark cheating for example? Nobody.

 

[teejee]

A Few watts is the max comsumption of a laptop medium sized screen.

What? You were wrong with a factor of 1000, and now you try educate me regarding the same subject?
I'm still correct, changing brightness affect power with several watts on a laptop. It is important to understand when interpreting test results.

 

[AtenRa]

The tests could be made with an external monitor, best option to eliminate laptop monitor dependencies

 

[Kallogan]

those seems to be packed with fx-8800P, full hd apparently :

http://www.saturn.de/webapp/wcs/sto...nel=sedede&searchParams=&path=&query=FX-8800P

 

[jime1]

those seems to be packed with fx-8800P, full hd apparently :

http://www.saturn.de/webapp/wcs/sto...nel=sedede&searchParams=&path=&query=FX-8800P

Nice, so its for real at last
tysm

 

[Dresdenboy]

The tests could be made with an external monitor, best option to eliminate laptop monitor dependencies

This is for us CPU geeks. Notebook buyers have different priorities.

This type of replies is exactly what comes to mind when others poop in Intel related threads. In the end all that remains is a strange odor all over the forum. What strikes me the most is many of those who enjoy taunting like this are actually rather educated, with solid technical background, and fully understand the forum rules and more importantly the clear benefits of abiding by those rules.

I get that part of the motivation behind such replies is the person debating with, and certainly Abwx has a very... abrasive posting style, but you should always remember for every person participating in the conversation there are many more listening, most of which deserve better than thread crapping.

+1

 

[monstercameron]

Carrizo-l is a rebrand of beema.