Does someone have an ELI5 on what AMD has done since ~2015 to turn their whole operation around like this, and why these new processors are so good? What's the catch?
Jim Keller was certainly key in getting Zen up & going, but he left in 2015 so his impact on Zen 3 would be rather minimal to say the least.
Combine that with the timing of Intel struggling hard with new process nodes and there you go. A bit of AMD got better, a bit of Intel got worse, and now you've got this landslide result.
AMD's execution looks meticulous by comparison, and they deserve all the praise and profits resulting from it. But their win is inflated by Intel completely imploding. Until next year they're merely iterating on 2015 technology - their Skylake architecture - on the desktop and in servers at least. This is not to say that Intel's 2021 chips will be competitive with this (they won't be) but if Intel with better management would have been at this point in, say, 2018 instead of 2021 this would have played out very differently.
So before they faceplanted with 10nm Intel first stumbled with 14nm. 2 generations in a row of issues.
Seems like a classic textbook case of a large coorporation not understanding how to change their ways, ending up being beaten by a more agile smaller competitor. Oh well, hindsight is 20/20.
Although AMDs and Dr. Lisa Su's achievement are not to be underestimated, I doubt the 7nm processes would be as mature as it is right now without Apple.
Apple's extra cash certainly didn't hurt, but TSMC wasn't struggling before Apple came along, either.
And prior to 7nm the other major fab, Global Foundries, as perfectly competitive with TSMC despite not having any Apple money.
Clearly TSMC is heavily diversified and gets to pick their clients. Apple has also been the primary (volume) launch partner on both 7nm and 5nm, which to me indicates how much TSMC values that partnership. Imagine the slam dunk Nvidia would have had if the RTX 3000 series was on 5nm.
They wanted more margins? But note that Nvidia does still use TSMC's 7nm for their largest & most expensive dies. The A100 is TSMC 7nm at a staggering 54 billion transistors on 826mm² of silicon. Nvidia retains the largest die manufactured on TSMC's 7nm. By a lot. The next largest would I think be Navi 21 at 536 mm² and 26.8 billion transistors.
> Apple has also been the primary (volume) launch partner on both 7nm and 5nm
Apple's die sizes & transistor counts are comparatively tiny. The first use of a new fab is pretty much always a small, low-power die. That's what yields best when yields are the lowest.
See also why Snapdragons are also among the first to launch on a new TSMC node, despite those SoCs having probably the lowest margins of the high-volume stuff TSMC makes.
Apple could have made the A14 much bigger and faster than the A13, but they chose to cram more chips on the wafer, which I speculate they did to have more wafers available for their bigger upcoming iPad and Mac chips.
I think you have it backwards on margins. Per square mm it’s clearly more expensive to produce an A100 die than an A13 or Snapdragon SoC due to the wasted die space and need for golden samples on the Nvidia side of things where they’re not selling any cut down dies in GeForce cards.
TSMC sells wafers, not functioning dies. They don't care how you slice it up or how bad you yield as a result.
It’s rumoured that Nvidia’s deal with Samsung was for working dies, not total wafers.
Highly doubt that he did so in a technical role, maybe in a PM like role, but not a design lead. The team lead was Suzanne Plummer, architect Michael Clark.
I was saying many times, if you hire a JK level cadre, it's best to not spend his very expensive time just doing technical work.
Let's just go back to 2014/15. And Look at Intel's Roadmap ( Or Slides ).
2016 - 10nm - CannonLake - First Gen 10nm Process. Industry leading Density.
2017 - 10nm+ - Icelake - New Architecture.
2018 - 10nm++ - TigerLake - Optimisation[1].
2019 - 7nm - Sapphire Rapids [2]
Most people would expect the ETA year to be +1. As 14nm itself was delayed.
TigerLake has around 15%+ IPC. That would put it ahead of Zen 3. The Intel 7nm is roughly between TSMC 5nm and 3nm. Had Intel shipped that in 19/20, along with Sapphire Rapid. AMD wouldn't even lead, they could barely compete.
So to answer your question. It wasn't AMD did some magic. They have been executing, and shipping their product on schedule. ( Although in tech, that in itself is pretty damn impressive ). All while Intel was.... um..... I dont know what they were doing.
10m fiasco has lots of other knock on effect as well. But I guess this is good enough for a ELI5 post :)
[1] This Process - uArch - Optimisation name came in later part. In 2014/15 people were still talking about Tick Tock.
[2] Sapphire Rapid was always a Server Specific design on the roadmap. But things changes bit. Along with Intel's plan of AVX512
Intel's next generation cpu node was supposed to come out this year, and it's been pushed back (possibly due to covid) until the end of 2021 or even into early 2022. Meanwhile, AMD says they have 5nm around the corner.
Unlike Intel who have micro improvements, each generation has been a big leap in performance so it’s been exciting for consumers.
Intel did something similar with their core2duos when AMD was on top around 12-13 years ago.
Those processors were such a leap over the Athlon XP 3200+ I had before and it runs a Linux desktop perfectly fine.
What are you using it for? Floppy disks were already very dead by 2008, let alone 2020.
I know there are alternatives to using a floppy to transfer to these ancient systems but for me it works okay.
For any practical purposes, Intel is now behind.
His previous stint on AMD was when he invented AMD64.
I think the 2 key high level choices AMD made was betting on TSMC, and betting on the chiplet + memory controller design.
A big part of his success story, was him having this great opportunity to develop his skills.
Microelectronics is a career suicide in comparison to software today.
I'd say 100 to 1 sieve off rate for people getting into the industry, to ones who actually get to do serious architecture work would be too small. The ratio is likely to be much larger.
We only can guess what they can be?
I suspect he would have left no matter what though, because Intel is so arrogant they wouldn't listen to him in the manner in which they would to.
That was the great thing about AMD; when you're already at rock bottom, you're humbled enough to listen to people.
AMD either did well or went broke which forces management to take painful but important decisions.
source?