Yeah because you are not a professional,in the industry giving your competition a 30% advantage is very bad for business,hey just look at the example at hand, AMD is forced to sell at 1/3 the price because the competition is (at least) 30% ahead...
Selective reading? My use cases for 8-16 core are completely professional and work related only.
I work in the industry so know quite a lot about it. I assume you do as well? Were talking about desktop CPUs btw and the 9900k is a CPU tuned for gaming not professional workloads. Professional workloads are about multi-core performance not single threaded performance where there 9900k shines. Where do you see 5Ghz :
https://www.intel.com/content/www/us/en/products/processors/xeon/scalable/platinum-processors.html
Nowhere because that's not the primary focus of
professional workloads. Moar cores is because the software is written carefully and with pure scalability in mind.
AMD is selling at a price that reflects they want to capture market share. Intel is selling at a price that reflects they still believe they wont lose any based on their legacy branding. That's a recipe for a reality check in the industry. It happens all the time. Ask various infrastructure providers what happened when the big data corps went white box wares and abandoned their ridiculously priced hardware some years back... a trend that is still not complete. Facebook's mission for instance ends around 2020 for forklifting external hardware out of their data centers and rolling their own. ARM is making strides. RiSc-V is on the horizon and already has penetration. Storage is going towards direct network attachment w/ no CPU involved via smart Nics and nvme. And here you are making false conjectures about AMD's pricing. The real pain hasn't even begun. Why do you think their upcoming chip is called 'Rome'? Have you even considered the I/O implications on that thing? That is my next likely CPU purchase. The majority of my 8/16 core rigs will be sold at that point. I bought into a platform and roadmap as most professionals do not some one off last minute attempt to seize the performance crown on a dead socket and process node.
As I consumer I see : price/performance value and roadmaps. Intel is unequivocally dead to me via this analysis. 30% more 8 core performance for triple or quadruple the price = DEAD... I can buy a 16 core from the competitor for intel's flagship 8 core price and did w/ confidence knowing there was nothing coming from Intel.. Wont be for years. I haven't even highlighted the fundamental security flaws in their processors.
Anand seems to disagree with you.
Number crunching = export to GPU
Video decoding/encoding = GPU
Offline Ray tracing = Now accelerated by GPUs
Video gaming = leisure hobby of wasting time (No need to go crazy)
Zipping/unzipping files = blink of an eye even on a 4 core and rarely done
Computing is going towards HSA not CPU centricity. CPU centricity was yester year's paradigm.
GPUs are capable of accelerating anything that is straight forward dumb but exhausting computation.
As such, CPUs are generally sufficiently clocked and performant as long as you have enough cores to handle a heavily diverse range of process execution simultaneously.
Intel failed in seizing the embedded systems/mobile market because they're intel.
Now Intel faces serious competition on both the desktop and server market.. because they're intel.
In the coming years, ARM and Risc-V will present even more competition as will the ever evolving architectures that put ARM cores closer to I/O and nix the CPU from the picture directly :
The emperor has no clothes anymore. Everyone now sees beyond Intel.
It's been a good ride up to 4 core with Intel. 14nm and pcie 3.0 is yester-year.
The future is 7nm and and PCIe 4.0 (AMD btw is rumored to have infinity fabric exposed over PCIE 4.0).
It's over my friend. Again, platforms and roadmaps vs performance crowns. Something professionals focus on.
Yester-year's company vs tomorrow's.