Quote from: _MT_ on September 05, 2020, 10:46:37

Quote from: john04092020 on September 04, 2020, 09:04:15
Who cares how it will perform. It is Intel so laptops using it will still score high in your reviews. No matter how it will perform, people will buy it because your review, even with half a dozen negatives, will still have a final score close to 86-90%.

If you look at the score, performance makes a very small part of it. So, it's hardly surprising. And IIRC, even then it's only gaming performance. Frankly, if someone can't be bothered to actually read a review and look at the data presented, I can't feel sorry for them. They fully deserve the stupid decisions they make. Even a child should understand that a rating obscures detail. That's, really, its job. Devil is always in the detail. And that "goodness" is subjective. A rating, no matter how good, will always capture a certain perspective (a set of priorities). It's a fact of life that Ice Lake often ended up in nicer laptops than Renoir. And you, in many cases, had a choice to make. Compromise on the CPU and get the laptop you want. Or the other way around. Or boycott the manufacturers.

No matter if someone's money go to a correct or a stupid decision, they will end up in someone's pockets. And in this case Intel's pockets. And that's the idea.

You can have 20 articles every day praising AMD's Renoir and just 2 reviews. One for an AMD model that scores 85% and one for an Intel model, at the same category, with worst performance and higher price that scores 88%. People will buy the Intel product. AND THAT'S THE IDEA.

Isn't using reference design models a standard practice across the industry? I have never seen Qualcomm use a model say.. Samsung galaxy s10 to demonstrate there performance. Or does AMD do that? Using reference models is a way the cpu manufacture can normalize the difference in the different end product. Products from Lenovo are different from hp are different from Dell or Samsung. Why should Intel or anyone pick one particular device to demonstrate. You as a review have that responsibility to compare the end products.

Quote from: john04092020 on September 04, 2020, 09:04:15
Who cares how it will perform. It is Intel so laptops using it will still score high in your reviews. No matter how it will perform, people will buy it because your review, even with half a dozen negatives, will still have a final score close to 86-90%.

If you look at the score, performance makes a very small part of it. So, it's hardly surprising. And IIRC, even then it's only gaming performance. Frankly, if someone can't be bothered to actually read a review and look at the data presented, I can't feel sorry for them. They fully deserve the stupid decisions they make. Even a child should understand that a rating obscures detail. That's, really, its job. Devil is always in the detail. And that "goodness" is subjective. A rating, no matter how good, will always capture a certain perspective (a set of priorities). It's a fact of life that Ice Lake often ended up in nicer laptops than Renoir. And you, in many cases, had a choice to make. Compromise on the CPU and get the laptop you want. Or the other way around. Or boycott the manufacturers.

That's why intel gives away their reference board design...

I agree with the point you are making here and history proves it. That said do you think the Evo brand on an equivalent config might mean they have to achieve similar benchmarks? Have they now set their own yardstick to jump over for any equivalent Evo certified laptop?

I love how Intel talked more about AMD than they talked about Intel in their presentation. Which processor is more relevant to consumers now, hint, its not Intel.

Who cares how it will perform. It is Intel so laptops using it will still score high in your reviews. No matter how it will perform, people will buy it because your review, even with half a dozen negatives, will still have a final score close to 86-90%.

These slides are more for the analysts, to convince them that you are on the 'right track' and they should continue to give you a 'buy rating'. What would you have them say? "We lost our way on Wafer Fab's and we will never get back on track now." "AMD designs better than we do now", "A year ago we messed up the supply chain and didn't have enough chips to sell." The truth does not sell well at all.
A decade ago I worked for a company where the president kept going to tech conferences and telling the analysts how our 'Nano Technology Solutions' were cutting edge and poised to take over the market, that is what the analysis wanted to hear, 'Nano-Anything' was the buzz word. Trouble is we weren't actually working on ANY 'Nano Solutions' at the time. No matter he kept the stock price up enough to cash out at a peak and sold the company off.
5G is the latest Buzzword, I am frankly shocked that Intel said nothing about their '5G Solutions' That is what the analysis want to hear now.

Quote from: Jim from Europe on September 04, 2020, 01:15:04
A is obviously an illiterate twit and technology neophyte. Probably an AMD fanboy...

"coprossesor"

AVX-512 is an instruction set extension available on Tiger Lake and other Intel CPUs and it not a "coprossesor"

Do you really have to resort to personal insults?

AVX512 was introduced in the Intel xhi coprocessor. But the point I was making is that it's not as simple as simply having your SHA accelerated instantly due to AVX512. You need to make sure your software supports that feature and that it is properly optimized. And even if you ask your software vendor to support it, they might not be able to until the library they use adds it or they transition to a new version of said library.

So maybe next time instead of hurling insults, try to actually comprehend what someone is trying to say.

A is obviously an illiterate twit and technology neophyte. Probably an AMD fanboy...

"coprossesor"

AVX-512 is an instruction set extension available on Tiger Lake and other Intel CPUs and it not a "coprossesor"

There are multiple issues that are at stake, thermals, how much wattage is used and etc.

But also because manufacturers like to cheat benchmarks a little like for example if they add a new coprossesor like AVX512. And show you that 1 app that happens to use it.

But the real question is in the GPU, since it is a new architecture, hardly any developer would optimize for it for at least a year or 2 if not more.