Partially. There are a couple of problems.
While Intel is free to make up whatever looney toon product labeling scheme they want (and they have done some hocus pocus scheme switches in the past ), a line up where everything begins with 32xx is probably wrong. Two factors. The current "gap filler"
Xeon W 3175X (with 28 cores and 255W ) now is to be approximately superseded by something numbered 3275 (no X so the clock rate is way down and probably not overclockable) . This is relatively consistent with Intel's processor naming for product range too.
X Y ZZ i
X -- SKU level in the product category ( in SP's Platinum , Gold , etc. for W at the moment this is proxy for Socket 3 - 3647 2 - 2011 and price level )
Y -- generation
ZZ -- processorSKU ( cores )
i -- processor options: some optional, 'value add' feature tweaks
So with the current scheme that would indicate switch sockets all the way down. That is somewhat unlikely.
Not only the socket is different but the PCH chipset
C422 ( for 2011 ) versus
C620 Series ( for 3647 ) . Those aren't the same cost, power , or board space ( CPU socket + PCH ) .
So those numbers would make more sense as
W-3275
W-3265
W-3255
W-2235
W-2225
W-2223
The > 200W TDP being the larger package size. Intel using
1P ( one socket) limited SP derivatives ( 'U' option) as Xeon W badged offerings (with bigger base clocks ; probably binned dies of the dual socket flipped off. ).
The second major "problem" is the dumping of the 18 core for a 16 core option in the W line up? In fact, the whole core count thing seems a bit missed up. The missing 10 ( max for the LLC , 'low core count', die ) is another head scratcher. Possibly Intel is going to do a another somewhat overlapping 22xx series that does up to 18 cores and are pushing here that the alternative, bigger W socket all the way down into the LLC die zone ( just to be cheaper with those incompatible motherboards ) and there would be some W-3223 with huge socket and bottom out at 8 cores. [ IMHO that's a waste of time, but there are large overlapping chunks of Intel's processor offerings that also seem to be an unnecessarily complex waste of time also. AMD's Threadripper is going to "crush" in the core count space not clock space. It could be a "shoot shotgun high number of SKUs at the side of the barn" approach to competing with AMD. I somewhat doubt Apple will buy into any of that. ]
As far as the leak in the Tom's hardware article I think that at least is a subset of the Xeon W offerings coming ( perhaps it is the initial shipping set. ). Things should be clearer at the end of the month.
There is a chance if Intel is going to "fork" the Xeon W line up into relatively complete "big socket" and "medium socket" line ups that Apple could go with the "big socket". The condition would be that they were not trying to go with the "literal desktop" Mac pro path. That they'd be shooting for a volume closer to the 2009-2012 models and not so limited desktop footprint targeted. ( that doesn't line up with rumblings, but it is all super secret ... so maybe). Still would be a one CPU socket workstation. They could go to 6 DIMMs without having two DIMM sockets per controller.
The downsides to that is that it caters to the folks who are out to dramatically increase the base cost of a Mac Pro significantly higher. The "big socket" 8 core probably will cost more than the "medium" variant (and a lower base clock). Filling 6 DIMMs with 8GB ECC will cost more than 4 (and with Apple's non market memory pricing that will have impact). There are several mechanisms for Apple to leverage to simply just push average selling price higher to the detriment of a decent chunk of the Mac Pro potential market and becoming a 'hobby project' on a pricing death spiral.
Apple staying on the W-22xx path or switching to AMD Threadripper would be a saner move if they really want to invest in keeping the product over the longer term as a non hobby project.
P.S. We'll see at the end of the month, but Intel could be doing a "Crazy Ivan" with the Xeon-W line up and pushing the 22xx sequence (with 2011 socket) out much further into the year than the 32xx update. AMD's Threadripper 3 may be slipping. ( perhaps because unexpected Eypc wins are consuming Threadripper wafer starts and AMDs is going after the higher profit margin. ). Also somewhat offseting SP package sales declines by pushing more of them in the Xeon W class where the competition isn't as steep. Intel would be throwing most of the 2011 socket packages at desktop Core i9 SKUs until the 14nm product cleaned up further later in Fall (and TSMC 7nm queue is less clogged with Apple's A13 ramp and AMD can change their wafer start queue mix better. )
[doublepost=1557674867][/doublepost]
I believe if apple actually got feedback from pro users (as they claimed as the reason for a such blatant delay) the mMP should be closer to the cheese grater design than my lived but flawed trash can.
Apple never claimed anywhere ( nor is likely to claim in the future), that the external feedback mechanism in the primary should of the delay.
The notion Apple couldn't start until the more formalized "Pro workflow group" was going and had been in opertiion for an extended time is a crock . Also that Apple go zero feedback from 2014-2016 is another one.
Days ago I wrote about a rumour/ leak that talks about a mMP being much like a cheese grater turned 90degress to keep vertical airflow,
turning 90 degrees is way past looney toons. First, the "cheese grater" is no primarily designed to be a literal desktop system. So the "rotation" shift thing makes little sense off the desktop. If you want to mount in a rack, 90 off vertical rotation axis isn't the only option nor does it have to be the primary rotation axis. if deskside mounted "more' fans sucking air off the dirty ground isn't an improvement for a computer system ( vacuum cleaner perhaps, but next Mac pro has no vacuum duties. )
Furthermore, if Apple is still manically trying to stick with the literal desktop shift for the Mac Pro .... that occupies a dramatically larger desktop footprint space. Extremely unlikey they would go higher than the 2009-2012 footprint which was already in the "oversized' to be a desktop footprint category.
I consider it much more plausible than the stacked mac-mini like mMP or the NeXT cube reborn, let's see, I'm ready to be disappointed again.
In a contest of assembling a large collection of less plausible solutions, the rank ordering of implausibility isn't much of an overall informative value add.
