I'll proceed to give the dead horse a beating, but what the hell is up with Intel's obsession with lakes? What was wrong with their old naming of architectures around a theme? Random landmarks, Dales, Fields, Bridges, Wells....
Why do they need to throw everything into a lake? Did Krzanich have a lake fetish or something?
Will the new CEO please give whoever is in charge of naming a beating and force them to come up with more distinctive names? Please?
They're following the same naming convention: every time they make a new iteration of the Core architecture (meaning: increase in IPC clock-for-clock with normalised RAM speed and timings), they switch to a new kind of landmark. Initially:
Bridges: Sandy Bridge (tock, architecture) ; Ivy Bridge (tick, new process node 22nm)
Wells: Haswell (tock, architecture) ; Broadwell* (tick, new process node 14nm)
Lakes: Skylake (tock, architecture) ; Cannon Lake*** (tick, new process node 10nm MIA)
Their fabs just screwed up big time with 10nm so Marketing Dept. took over to save the day and they came up with the
Hyperscaling and
Tick-Tock-Optimization shenanigans. Thus, the "Lakes" generation is adding a new member to the family for every year they fail to deliver 10nm:
Lakes: 2015Skylake (tock, architecture) ; 2016Kaby Lake (
optimization,
more Mhz) ; 2017Coffee Lake (
optimization, moar cores) ; 2018Coffee Lake Refresh? (
optimization, even moar cores)
Given the struggle they've been through with 14nm (Broadwell's launch on desktop was a joke) and given that they forecasted that 10nm will have worse transistor performance than the refined 14nm, they decided to iterate on the "optimization" thingy and thus 10nm will have no
tick or new architecture until the process gets refined.
I'm guessing that adding AVX512 in consumer cores could be seen by some people as a
tick, but given that Intel themselves are naming the new families Cannon Lake, Ice Lake and Tiger Lake, I wouldn't call them new architectures. So the Core family has a ton of lakes just because Intel's struggle with 14nm and 10nm:
Bridges:
Sandy Bridge (tock, architecture) ;
Ivy Bridge (tick, new process node 22nm)
Wells:
Haswell (tock, architecture) ;
Broadwell* (tick, new process node 14nm)
Lakes: 2015
Skylake (tock, architecture) ; 2016
Kaby Lake (
optimization 14+) ; 2017
Coffee Lake (
optimization 14++) ; 2018
Coffee Lake Refresh? (
optimization) +
Cannon Lake-U* (tick, new joke of a process node 10nm) ; 2019
Ice Lake (optimization, 10+) ; 2020
Tiger Lake (optimization, 10++)
Rapids:
Sapphire Rapids, initially planned for 2020 (tock, new architecture 10++) ;
Granite Rapids, initially planned for 2021 (tick, new process node 7nm presumably EUV)
No wonder why you're tired of Lakes

. But as far as I'm concerned they're actually following the same naming scheme for 10 years, it's just fabs are a huge mess now so they're stuck in the same landmark until 10++ gets sorted out.