It is very important to realize these for mobile chips and will likely i5 and higher performance wise. Also since the chips have more cache and Xe graphics they will be 11 or higher series. The level you are referring to likely Lakefield class of processors with hybrid single Sunny Core with 4 new Tremont Atom. Keep in mind Tiger Lake as AVX-512 which is similar to what is on Xeons now
sigh
The i# tells you where the process sits,
within a tier, but doesn't tell you what tier it's in. I would hope forum regulars picked up on that, by now, but I know some people don't pay much attention to mobile. If you want to know what tier it's in, look at the series: Y, U, or H (the segment which used to be QM).
This is a U-series, meaning it's mainstream mobile. In the next couple years, that's going to be 10 nm, which have better power efficiency but generally worse performance than 14 nm (hence, why we just saw Comet Lake launch into mobile @ 14 nm).
Whenever there's a juicy roadmap leak, that's something
you might want to bookmark:
https://www.tomshardware.com/news/intel-roadmap-10nm-14nm-gpu-cpu,39163.html
One note about that: it's for commercial laptops (read: business), so the dates will lag the offering of those CPUs for consumers. However, the names, tiers, and technologies shouldn't change.
The inclusion of AVX-512 is just a standard trickle-down of high-end features into the mainstream. I think it'll be of limited usefulness, but it's mainly aimed at AI/inferencing (which really belongs on a GPU or specialized hardware block, but that's a different discussion).
Also if you closely at benchmark article included in the article - it compare the Y version - you know what Intel use to Core-M and compare it competition
It was a different article about a different leak, relating to a different (Y-series) Tiger Lake CPU. The results are therefore of limited relevance, but what you omitted is that Y-series CPU's performance relative to a Coffee Lake U-series:
When pitched against the i7-8559U quad-core
Coffee Lake processor, the Tiger Lake Y chip seemingly performs just
4%, 2% and 8% slower in single-core, quad-core and multi-core workloads, respectively.
Here's the specs on that i7 U-series (also 4C/8T; 20 - 28 W):
https://ark.intel.com/content/www/u...-8559u-processor-8m-cache-up-to-4-50-ghz.html
However, the caveat is that the benchmark used is probably not the best CPU metric. I'm not familiar with UserBenchmark, but it sounds to me like one of those "user experience" benchmarks that are heavily influenced by more aspects of the system, such as GPU and storage. And we don't know what sort of system that i7-8559U was in, but it was likely a desktop mini system that could've had a SATA SSD (or slower NVMe). So, I really wouldn't pay much attention to it.