....
Until Intel has concrete dates for Xeon chips, it's pointless to get worked up much.
Intel has said Cascade Lake ( Xeon SP) was to be Q4 2018. ( SC 18 Supercomputer Conference is Nov 11-16 probably will get more clarity at that point whether that has slid how much into 2019 . ) That is probably going to slide a bit due to the 14nm log jam for "normal" vendors. ( Google , Amazon, MS Azure , some SC vendors will get theirs first. )
Normally the -W ( which uses a variant of some of those dies and substantially different packaging) would follow within a month or two. That appears to be a bit muddled. Probably because of Threadripper 2 timing, Intel is doing a "Skylake-X Refresh " for the HEDT variant of the Intel -W's baseline.
https://www.anandtech.com/show/13402/intel-basin-falls-refresh-core-i9-9980xe
No meltdown/spectre fixes. They are just partially removing the PCI-e lane kneecapping they did to desktop line up (still don't get full 48 ) and activating the caches (even without associated cores ). And keeping the pricing the same. I'm not sure how that will 'save' them competitive sag much at the top end of the mix, but it is probably something they can limp on a very mature process/die they have been doing for a long time.
Intel may skip doing that weak 'Refresh' with Xeon W. The PCI-e lanes already were not kneecapped so there is
absolutely nothing to 'refresh' there. A 'refresh' that only amounts to some incremental L3 space on the lower half of the product line up is extremely marginally useful. Having no Meltdown/spectre fixes means still taking the same perf hit. Someone at Intel would be smoking lots of something if they thought that was really going to make a difference competing with TR2.
I suspect they will just limp along with W in its current state until late Feb- early March (when the initial demand bubble for Xeon SP dies down a bit) and then do W with the Cascade baseline ( fixes and maybe more cache at the lower end. ). Intel's $1B is suppose to get more 14nm capacity online around March-April 2019, so they could move forward at that point. ( and in late April - June dump this relatively lame desktop X "refresh" too. )
Intel may be desperate though and do a super lame Xeon W "refresh". ( AMD is going to lob some hand grenades towards Intel's direction next week. (
https://www.anandtech.com/show/13538/amd-investor-relations-next-horizon-november-6th ) Post the SC 19 conference it all should be move clear (because AMD and Intel will be lots of "We're better than them" talking around that time and he Workstation stuff should indirectly fall out of all that. ) .
At this point I would expect WWDC though.
WWDC to ship perhaps, but I'd expect them to do/say something before that. If Apple is almost ready but the Intel part slide then they'd wait until just how far the slide was going to be to until start talking.
Even if Apple did something completely bonehead and only had a partially complete system at this point so that they are targeting late 2019, they'd still need to cough up a much fine tuned timing window around April 2019. The substantial problem is going to be even if Apple doesn't uptake on those Q1- earlyQ2 new Intel W's, their competitors are. The iMac Pro also appearing to start to go stale is just going to add to the smell coming of the MP 2013. If their now annual "dog ate my homework" meeting happens to approximately match up with the competitors actually doing something, that is worse than just the "dog ate my homework" meeting by itself.
[ The new Mini is beating the general iMacs on CPU heavily workloads. It is hard to believe that Apple is going to 'coast' on that for another 7 months in another "can't say anything until WWDC" move. There is no good reason for Apple to throw away the first half of 2019 playing dumb. ]
Also, given the moves to the mini, I'm curious what the iMac line will look like. I can't see them going all-flash unless they update the chassis at the same time, but at the same time I'd love for them to finally leave spinning boot drives behind even if it comes with a dumb price increase.
Apple doesn't need to update the iMac chassis to go all Flash. They could just do it ( and just not repurpose the saved 3.5" space for anything. ). Backsliding from a two drive system to a one drive system doesn't need much but a logic board change ( to remove the 2nd drive connector).
That wouldn't get them much better thermal cooling capacity, but they could switch.
The price increase isn't 'dumb'. There are more than a few Mini folks (for the lower end of old price spectrum) who are not very happy at point.