Olle P :
Since RTX-cards make up for such a small amount of the market (user base) I can't see much profit in spending lots of time adding RT functionality and optimization to a game. Having some RT is good for marketing, but on top of that it will cost more than the sales will increase.
Yeah, I doubt that any big games will require hardware raytracing support even two years from now, especially if the mid-range 20-series cards lack support for it, which seems likely. And even most of the games that support the feature over the next couple years will likely only implement it as an afterthought, since the vast majority of those playing the game won't be able to see the effects. They'll focus on polishing the games to look great on the current-gen consoles and video cards that almost all players will be using, and as such, less time will likely be spent on raytraced effects. Nvidia may pay some developers to implement RTX in their games to act as tech-demos for their new hardware, but I can't see it becoming the norm without a big install-base. And if the cost of entry for raytracing-capable hardware is $500+, then there won't be much an an install-base for quite a while.
brunis :
I think i'll take the wait and see attitude. AMD is about to launch 7nm Zen2 cpu and new GPU's based on the same process. According to Samsung they got 20% perf and 40% lower power from just 10nm -> 7nm. AMD is jumping from 14nm -> 7nm .. so, a loose guestimate, without cpu arch changes or improvements. 40% more power at 80% less power.. Too good to believe.
Too good to believe, because those numbers are quite unlikely. : P And I believe Samsung was referring to their process offering more performance
or lower power, not both at the same time.
I'm also not expecting the next generation of Ryzen desktop processors for about half a year. The Ryzen 1000-series launched around March and April of last year, the 2000-series in April of this year, and my best guess would be that the 3000-series might launch around the same time next year, assuming that they are getting good yields from the 7nm process by then.
jimmysmitty :
More than likely AMD will see a 10% +/- performance increase mainly from higher clock speeds at the same power or less power at current performance numbers. Same has gone for Intel in most generations as the new generations typically can increase the base clock while staying in the same thermal envelope or lower that thermal envelope while maintaining performance.
Recent rumors have suggested that there could be around a 13% boost in IPC over Zen+, along higher clocks, so if those are to be believed, then performance gains of around 15% or more could potentially be possible, at least at some tasks. At the very least, I wouldn't be surprised if per-core performance was right on par with Intel's processors, if not slightly better, at least until Intel's 10nm CPUs come out, whenever that might be.