It was only partly AMD, mostly TSMC. 3D fabric is a TSMC innovation, and if AMD chips consumed considerably more power the non 3D just wouldn't be able to compete, much less win. Raptor Lake is still DUV lithography. AMD did not make EUV, they ordered it from TSMC, like 3D cache. The 9950X would not look very good compared to the 13900k if it were on 10++.
somehow you turned intel failings into AMD mediocrity. amazing blue goggles, do you work for Userbenchmark?
Here is the facts of the matter. In 2011 with complete domination in hand in the pc market and the new smartphone/tablet market opening up Intel took a gamble. they were going to prioritize mobile cpus, with the end goal of shrinking their x86/x64 leading Core tech down to mobile sizes and voltages, bringing the desktop to the tablet, with an eye toward dominating the mobile market as well. This shift in design philosophy emphasized node shrinkage, and power savings over performance gains. they also split off a large chunk of their engineering department to work on the Atom processor as a stopgap to get their foot in the door at the mobile level while they attempted to transition desktop power to mobile form factor.
-note, intel was not interested in the desktop market that they now owned, it was not their priority. and enter an era of 5%-10% performance gains largely from efficiency gains and node shrinks.
well Atom failed, hundreds of thousands of manhours and millions of dollars of money burnt, and the node shrinking hit a hard technical wall intel couldn't overcome...
Intel, still with no real competition in desktop chose to attack the node shrink issues and pored all their R&D into fixing their problems getting past 10nm, meanwhile Intel had hit the limits of their Core cpu structure with Skylake. they relied on lithography improvements and increased clock speeds to stretch out the lifespan of the chip design, rather then design a new one, as all their assets and time and money were being pored into their node shrinkage... this looped them 6 generations of refreshes with no significant design changes while they milked the pc public.
What's fascinating is Intel had gone through a few CEO changes over this time, and it's largely believed they were still pushing the desktop on mobile design philosophy due to lack of direction from the top. as far as the top was concerned as long as money rolled in everything was good, so the design philosophy and direction of their R&D wasn't really touched...
Ryzen hit the market between the eyes, but Intel were slow to recognize the problem. if you all remember Ryzen one hit the market in 2017... during the 1st skylake refresh (7th gen). Intel still didn't wake up to the problem Ryzen represented, because at the time Ryzen, while competitive was still behind by about 5% in IPC (ryzen launched with very low clocks, and while it was about on par with intel at 4ghz, the clock speeds of 8th gen opened up a significant 15% or more lead in actual performance). Furthermore Intel had just started in on it's lithography issues and still believed it was a minor problem they could lick with a little more time.
While Ryzen went through it's growing pains intel stretched out the skylake architecture for 5 more refreshes pumping clocks and languishing on 10nm... lack of direction from the top and confused mission in engineering meant intel was mostly just playing with themselves while the house burnt down around them
They wouldn't wake up to the problem they had made for themselves until Ryzen 3 and x3d hit the markets. REALLY late in the game to realize they had lost hold of the market. But i think they could have survived this had they not completely dropped the ball on datacenter cpus. That was where AMD did the real damage to intel. While AMD/Intel cpus were a toss up durring ryzen 3, in datacenters it wasn't even a game anymore, with AMD server cpus vastly out performing intel xeon chips, by significant margins, in pretty much every metric, power draw, core count and raw performance. AMD's chiplet design really showed it's power in the datacenter market, and that was were AMD made it's fastest and most significant inroads and did the biggest damage to Intel's profits.
Intel's xeon architecture is really showing it's age, unlike the rest of the skylake refreshes intel couldn't just increase the clock speeds to make up for a lack of technical design improvements.
Facebook
Twitter
Instagram