i like the "media separated from graphics" bit a lot. this should significantly reduce youtube/netflix playback power consumption. its unreasonable that i have to use 10+ W just to stream 1080p videos

Large companies want access to large wholesale from the processor manufacturer. But if AMD is not able to fulfill the order at a given price (they are simply not interested in it, i.e. it is unprofitable in the absence of their own factories, since they themselves have to pay TSMC, a significant part of the margin), then make a line for AMD from a gadget manufacturer there is simply no desire, i.e. commercial interest. For M$ this is not a core activity at all - everything is done, most often, so that the employees of the relevant related departments can justify their existence and the costs of paying them. They, such non-core divisions, have no interest in making a truly massive and mega-popular line, which always appears only due to the optimal price (strictly tied to purchasing power in the form of a % in relation to the median income in each individual country , i.e. what is popular in the USA may be completely unpopular, for example, in China, due to the radical difference in median incomes) and a certain set of model parameters that best meet the aspirations of the mass buyer.

And what shapes the aspirations of the ordinary, mass buyer, given their almost complete lack of technical literacy and experience? Expert opinion. They are the ones who set real trends in the market. In the presence of influence on mass consciousness, which cunning marketers are trying to intercept - specialists in mass psychology and manipulation of consciousness, but never experts in anything else. It is simply impossible to explain all those failures in model design otherwise. When everything is obvious at first glance, but continues to remain in such a poor state for years. When all the companies conspire to turn the market into an oligopoly, the buyer simply has no choice. And there is no way to vote with your money and your feet in the other direction - because... The entry point for new players (with a comparable level of production quality and mass production) is monstrous in the modern world. And the ugly patent system, deliberately, carefully developed by systemically corrupt legislators, in all countries, over decades, for the benefit of TNCs, and not start-ups and small companies, completes the picture.

I would be very happy if 2-3 more large players appeared in the world both in the market for the production of equipment for the production of advanced chips, and 5-10 new large players in the final product market at each level of the chain further, but alas, this is only pipe dreams. While we only have a single ASML as an advanced silicon chip manufacturing equipment supplier, there are 2.5 advanced chip manufacturers. 2.5, because Intel is so far only boasting of a future "breakthrough" that by 2026 it will bypass TSMC/Samsung in technical processes.

On the other hand, TSMC also stalled with the introduction of "2nm", postponing its introduction to 2026. As expected by experts, productivity growth is gradually turning into an increasingly flat curve on the graphs.

And all this together obviously leads us to a dead end and stagnation in the growth rate of real productivity per 1W of consumption.

Namely, performance per 1W is the only measure of success in IT progress for civilization. After all, new classes of tasks that qualitatively change the capabilities of the individual and civilization as a whole require an increase in productivity not smoothly, by several times, but by several orders of magnitude at once and by orders of magnitude greater demands on RAM and long-term memory. At least in von Neumann architecture. At least in the same attempts to create advanced neural networks, which are deliberately advertised to the illiterate population as "AI". Neural networks will only become valuable when they produce effective solutions in your pocket, and not when connected to some global network. It is the "power supply" (in all its senses, from the literal to the computational, including the general level of intellectual development) of an individual that determines the overall real progress of technological civilization.

Civilization is actually now still reaping the benefits of the fundamental breakthroughs of the 20th century and the dramatic growth of the 20th century. In the 21st century, everything is still bad with new developments. And there are no real breakthroughs anywhere while humanity wanders in the dark, in its inability to formulate new ideas and means of their implementation. Even the crisis in fundamental science is becoming more and more obvious. In addition, the return from the frantic population growth over the past 100 years, in terms of accumulating return on human capital, is, as is already obvious, less and less, i.e. the efficiency of civilization is rapidly falling against the backdrop of problems with our habitat.

Quote from: Dan6 on September 20, 2023, 17:52:41

Quote from: NikoB on September 20, 2023, 13:42:21... Intel will not be able to offer anything in 2024 that would be really better in terms of energy efficiency than even the current Zen4 Phoenix/Zen4, and even more so the polished Zen5 2024.

They have a pile of cash to pay to laptop manufactures to ignore Ryzens (as well as licensing and other ways to manipulate). I want to buy some premium 13-14" AMD laptop without dedicated graphics and the only option I have is Thinkpad T14, which in my country is limited to 16gb of RAM with Ryzen, while being available with 32gb on Intel model.. (also to mention that AMD version has soldered RAM, while Intel's doesn't.. why?)

P.S. And here is Microsoft, releasing Surface laptops in couple of days with 13gen CPUs... Then they'll say that for some reason their laptops are not popular and close Surface department..

Surface laptops are not popular since Microsoft is reluctant to offer AMD CPU options to their product lineups

Quote from: NikoB on September 20, 2023, 13:42:21... Intel will not be able to offer anything in 2024 that would be really better in terms of energy efficiency than even the current Zen4 Phoenix/Zen4, and even more so the polished Zen5 2024.

They have a pile of cash to pay to laptop manufactures to ignore Ryzens (as well as licensing and other ways to manipulate). I want to buy some premium 13-14" AMD laptop without dedicated graphics and the only option I have is Thinkpad T14, which in my country is limited to 16gb of RAM with Ryzen, while being available with 32gb on Intel model.. (also to mention that AMD version has soldered RAM, while Intel's doesn't.. why?)

P.S. And here is Microsoft, releasing Surface laptops in couple of days with 13gen CPUs... Then they'll say that for some reason their laptops are not popular and close Surface department..

I mean, you can see it like that ("beta chip"), but you can also see it through a different lens: For the mobile market, efficiency matters more than for the desktop, because battery life, fan noise and heat emissions affect the user experience of a laptop. So it makes sense that Intel would put their most advanced chips into the laptop market first, while "stretching" their older platform a bit longer on the desktop.

Also, laptops are a far bigger market nowadays than desktops, so it also makes sense for Intel to prioritize the mobile market.

Quote from: NikoB on September 20, 2023, 13:42:21From the point of view of the average buyer, Intel will not be able to offer anything in 2024 that would be really better in terms of energy efficiency than even the current Zen4 Phoenix/Zen4, and even more so the polished Zen5 2024. But again, everything will come down to AMD's banal lack of production and its loss interest in the SoC market for laptops...

Yeah. The only thing that will count in the end is availability. And AMD doesn't seem to have interest in that. So if Intel can produce much more volume and higher availability it will be the only real option for the people, despite how good AMD APUs are theoretically.

QuoteIntel always has a lot of bugs in the first generation on a new technical process and it always becomes the most buggy for them. Everything is exactly according to "tick-tock" practice. We first make a "beta" version using a new technical process, test it on customers at their own expense, and then release a more polished version of the same thing under a new name.

Exactly. I didn't follow Intel developments that close in the past, but it's exactly like that...
- Release a beta chip for mobile only (Meteor Lake)
- Release more polished verson for both desktop and mobile (Arrow Lake)

And afterwards again:
- Release a beta chip for mobile only (Lunar Lake)
- Release more polished version for both desktop and mobile (Panther Lake)

While this scheme makes sense, it's not very motivating to buy a computer with beta chips. Unfortunately, after so many years with no improvements on the mobile section, and now coming up with something significantly better, it will be hard not to buy Meteor Lake chip. One would have to wait another 1.5 years for the polished version (Arrow Lake)...

From the point of view of the average buyer, Intel will not be able to offer anything in 2024 that would be really better in terms of energy efficiency than even the current Zen4 Phoenix/Zen4, and even more so the polished Zen5 2024. But again, everything will come down to AMD's banal lack of production and its loss interest in the SoC market for laptops, with further winning back the market share in favor of Intel, albeit with more backward "7nm", but producing SoCs for laptops in a ratio of at least 5:1 in their favor. As can be seen from the real shelves.

In addition, as real practice shows, Intel always has a lot of bugs in the first generation on a new technical process and it always becomes the most buggy for them. Everything is exactly according to "tick-tock" practice. We first make a "beta" version using a new technical process, test it on customers at their own expense, and then release a more polished version of the same thing under a new name. Business is going on...after all, buyers have nowhere to go in the Intel/AMD oligopoly in the x86 processor market, imposed on us 20 years ago with the connivance (intended for geopolitical purposes) of the American antitrust authorities...

This looks very promising. Especially, hearing 2X transistor-concentration and 2X iGPU perf/watt is incredible. I am happy to see Intel finally overcome lithography problems and properly adopt EUV lithography. Good job Intel! You are in the right path. Efficiency must be the goal 👍

It needs much greater bandwidth to be anything but trailing edge for emerging applications. Even the on-chip memory option only has 120gb/s.

No way this is not bot 😆

Intel has finally woken up after almost 4 years of manufacturing the same thing, now we will have to see the performance and energy savings of these new laptops and their new iGPU.
One thing I am clear about is that currently if I had to upgrade the laptop I would choose Zen 4 7040 Phoenix because it has demonstrated efficiency and performance and because I do not plan to spend more than $2,000 on any laptop. I am afraid of the prices of these new Intel compared to AMD, which despite not having factories, is doing a spectacular job and Zen 5 has not yet arrived, which is considered something BRUTAL for 2024