Which is better? (CPU)

Intel stated that 8th gen is not compatible with z270 because they modified some power pins in the cpu so z270 will never become compatible with 8th gen.

https://www.gamersnexus.net/hwreviews/2827-amd-r7-1700-review-amd-competes-with-its-1800x

Just a few game examples further down, so you'd want more examples (and Ryzen probably is doing a bit better now) but you can see how things compare... AotS is more of a "demo" to most people than a game. It's very atypical of most game CPU usage and feels like it's been kept around to showcase both the CPU and GPU AMD architectural changes.

I'm not sure the i5-7600K is far enough ahead overall that I'd stick with it, but your choice should be best bang-for-your-buck for the ENTIRE BUILD not just those two CPU's. Many factors including the fact you bought the DDR4 memory and motherboard already.

Unfortunately in the long run there's not much upgradeability to that CPU socket. The best I would assume is the i7-7700K which is too expensive to recommend when the i5-8600K has 50% more cores (6C/12T) and last I checked cost the same as the i7-7700K.

I'm not sure what the motherboard cost, or if it can be returned but if you want to open up a discussion that's fine.

The i5-7600K is a great CPU so perhaps I'm nitpicking. I'm just used to usually small upgrades to existing PC's (like graphics cards) or building a new system.

Sure.
It's not just the Frequency but also IPC (Instructions Per Clock) which are both higher on the i5-7600K.

(the IPC will change a bit with updates to software but no need to dig that deep)

Really, the i5-7600K is a great CPU so if you got the motherboard already let's just put this behind us and decide that's the best choice.

Photonboy please if you dont know then dont spread misinformation, IPC has nothing to do with software, it is purely hardware and it is the number of cycles can do a cpu per clock. What you say by software is the multithreading and that can be optimized by software. An example for such an update is Rise of the Tomb Raider which got a dx12 update for ryzen boosting performance by 20-30%. That was done by optimizing the game to use more threads more efficiently.

You are the one who is misinformed on IPC.
https://en.wikipedia.org/wiki/Instructions_per_cycle

"The number of instructions executed per clock is not a constant for a given processor; it depends on how the particular software being run interacts with the processor,..."

In fact, the COMPILER used to create the software is very relevant to optimize how much code the CPU can process given the same number of cycles. So a future game or application using a compiler optimized towards Ryzen's architecture should produce better results, thus its IPC will go up for that particular application.

This could be due to managing thread jumps, optimizing the buffering or whatever but SOFTWARE does matter. IPC is not just a simple number assigned to a particular CPU (though of course you can compare the same PROGRAM).

It's not that important for this post, but I hope you learned something.

I have this strange feeling you're going to argue the point but just stop and read the part I QUOTED that uses the word "software" which you said IPC had nothing to do with. If you really feel the need to argue then you can PM me.

(also you didn't state it correctly. You meant probably to say "instructions per cycle" not "number of cycles can do a CPU per clock" which doesn't make sense)

It seems we are both wrong, Wikipedia is known to have mistakes and this is one of them, watch this video untill the end https://www.youtube.com/watch?v=gLsdS0zQ82c .

It seems that in the first place is on hardware level and second is at software level, if the cpu has high IPC but the software doesnt do parallel computing well then the cpu is not fully used, if the software has high parallel computing the the cpu has low IPC then the cpu is the bottleneck. From what i see Ryzen has a Hardware limitation IPC and one proof is single thread cinebench. Cinebench has really really good parallel computing single and multithreaded so ryzen should perform great there if it had good IPC but doesnt do so good, with my 1700X i had max 165 points on single thread on R15 while on my broadwell i have 190, so i think ryzen has a low IPC and no matter what compiler is used it will not achieve intel IPC because it is at hardware level. Thanks to you i had to do more research now things are very clear to me and learned 2 things:
1) Software IPC how you called it is called parallel computing and how good is a software for parallel computing to unlock full potential of cpu IPC
2) IPC is exclusively at hardware level but if the software does not do parallel computing then the cpu has a high IPC for nothing because it will not be used and from here comes the "software IPC gains" you mentioned. But in reallity a CPU IPC cannot be increased because it is hardware limited by the "pipeline" mentioned in the video.

Saturnity,
For the purpose of my IPC comment I was referring to the ability of RYZEN to improve relative to Intel in the future. I don't mean a particular CPU would suddenly do better than a particular Intel CPU but just in GENERAL. For example, Intel might gain nothing on the same CPU but Ryzen might execute the same code 10% faster (out of my ass numbers).

Again, one way is to keep the same THREAD of code running on the same physical CORE of the CPU as there's a latency penalty for jumping between cores, and worse if you're jumping to another CCX (for Ryzen). So when a compiler is written such that the machine language is executed to reduce jumping and in general be efficient at not wasting clock cyles that's a good thing.

A program like Prime95 isn't very realistic but it does a great job of moving known chunks of data in and out of CPU's (Intel especially) and minimizing wasted cycles. That is also why these CPU's run much HOTTER than normal. (the same principal applies to GPU's too... if the game uses most of the usable transistors by good thread management and such the GPU consumes more power).

I'm not sure if there are ANY programs compiled with a Compiler optimized specifically towards Ryzen. That could make a difference. Just FYI, since you may not read this.

(Dragos Manea, I think you are mixing up multi-core/thread with the pipeline width. Specifically:

"Each execution unit is not a separate processor (or a core if the processor is a multi-core processor), but an execution resource within a single CPU such as an arithmetic logic unit."
https://en.wikipedia.org/wiki/Superscalar_processor

It's not a simple thing to understand actually, and there's little benefit to understanding much beyond benchmarking and how multi-core CPU's with or without hyperthreading work.

IPC as a set number (i.e. theoretical maximum) or variable value for the same CPU based on whatever software is used (i.e. benchmarking) seems to be done both ways so I think I'll just avoid the term and say something like:

"look at the performance for ALL the applications you use to determine what CPU to get... some will do better if the extra cores/threads overcome the lower performance per core, whereas some GAMES especially benefit most from a few fast cores.. "

I'm done though. CU.)

I was refering for out of order processing aka parallel decoding in a single core (the ipc you are talking about). If a core can do lets say 16 out of order decodes but the software provides only 10 then it is a waste of resources. Did you even watched the video? You keep pushing me that wikipedia crap in my face.

Honestly ryzen in gaming will ever be haunted by this ccx innovation, it was a mistake from amd to lunch a gaming product which got crippled from their own technologies. Yes it is a great value for budget adn evena greater value for video editing and 3d render machines but they could have done things better. At least to let us overclock to a 4.3-4.4 ghz to compensate for that ccx. This is coming from my personal experience, had a ryzen 1700X.