When betwixt and between generations there are several things to consider:
a) I always advise staying of the bleeding edge". The P67 is one of the nest examples of this whereby the original stepping were found to be defective. The problem was fixed in the B3 stepping. Even tho manufacturers offered free replacement boards... who wants to to a complete rebuild ? We always recommend waiting at last 3 months'; when i did a Z87 build, there were early the board I had chosen had a problem with after "going to sleep", when waking up, all external storage devices wouldn't wake up... buiuld wound up being delayed from August to September till the C0 stepping came out which eliminated the problem. In addition, as production lines are tweaked, yields and overclocking potential improves .... where as xx% might be stable at say 5.0 Ghz with the early CPUs, in later steppings, that % always improves.
b) Refresh or new design ? Not every generation comes with significant changes in technology or feature sets. For example if you were wondering whether to build a new box for a generatiion that was the last to support DDR3 ... or wait for the next one with DDR4 support, that's worth wating for. Other inprovements might include extra PCI lanes whci affect your storage choices or external connectivity.
c) When the technology change is merely a tweak of existing tech and offers speed changes, it's a matter of weighing the increase in cost versus the performance gained. That decision will be unique to each user.
d) As far as the notion that AMD has caught up.... this is both true and not true at the same time depending how you define "caught up". Each platform has it's strengths and weaknesses and does caught up means equalled or passed' or does it mean "close". We build PCs for a purpose and the selection should involve no other factor other than how it meets those intended functions. Things that do not matter are, die size, number of cores. Being smaller doesn't improve performance ... having more cores doesn't help if you are not going to use them. Performance in benchmarks or in things that you never do or do once a year or once in a PCs lifetime should not be involved in the decision process.
Here's an example .... a user came to us with the intent to do a home build for the following usages and couldn't decide which platform to choose. Had been involved in a few similar builds with this usage profile of late and and where in the past, Id just say "this is what you should chose" ... this time i just pointed to TPU reviews and suggested he do his own comparison. The primary usages were.
a) Gaming
b) Work from home (CAD)
c) Wanted to get into Photo Editing / Video Editing.
d) Office suites
I suggested he go to TPU site and look at the reviews looking at reviews for the 3900X and 9900KF. We had a subsequent discussion covering topics such as:
CAD - he noted that a CAD sites he had looked at recommended the 3900X cause it was significantly better in rendering 2D and 3D CAD drawings into 3D models. But he had no rendering software and his firm had never rendered drawings for any project and only used AutoCAD for 2D Drafting. For 2D and 3D, specifically with AutoCAD, it's an Intel / nvidia world. We also discussed what if his form decided to delve into rendering, I suggested that if ya have say 9 CAD workstations, adding a 10th 3900x based box as dedicated rendering station would be best.
Photo / Video Editing - He noted that various photo / video rendering sites had recommended the 3900x and others 9900K. We discussed that performance depends on the application being used. Since he's be using Adobe products, suggested he concentrate focus on those apps.
Gaming - this of course is challenging to pick a best as you could "prove' either outcome merely by pick the games used for testing. While this presented a pickle ... if ya don't wanna choose by looking at which games you play ... there's always TPs summary.
He eventually chose to go Intel and while he thought that Intel would perform better, the difference was not such that he couldn't be swayed by cost or other factors. The tipping point was the 20C CPU temperature 30 watt power difference (room didn't have AC) and wide community overclocking support.
Point being here... make your choices for the right reasons ... how fast a piece of hardware completes a task it will never see is not relevant.... die size is no relevant .... cores are not relevant unless you need more of them. Take the same approach to when to jump into an upgrade ... resist the urge to have a new shiny thing on yoiur desk, make the change only when the benefits are worth the investment.