I made another topic to further explain this phenomena, you should go read it after reading this one.
It's graphics card upgrading season and the one thing on a lot of people's minds is "will my system bottleneck this graphics card?" And the answer to that is... it depends. Not just on your hardware, but the game itself.
What is a bottleneck?
I'll leave it up to this video to explain it.
But if you are the type who wants me to get to the point: a bottleneck, or specifically, a CPU bottleneck, is when the CPU is too busy that it bogs down other parts of the system. With games, the CPU is too busy to tell the GPU to render stuff and you get lower performance than what the GPU should be capable of.
How can I tell if my system will bottleneck a video card upgrade?
The basic way to tell is if your CPU load is constantly high, you'll have a bottleneck. How high? I peg it at about at least 80%-85% almost all of the time you're playing the game. If you want to know if your system will bottleneck a potential upgrade, run the games that you usually play or want to play, fire up task manager (CTRL + Shift + Esc if you haven't learned the other three finger salute), and leave it on monitoring the CPU utilization window.
This is what my CPU looks like during a GTA V run
I say that it has to be at the 80%-85% almost all of the time because a game can be highly dynamic. If there's enough explosions going on to make Michael Bay shed a tear, then the CPU is certainly going to be hammered a lot. But often times even a regular ol' shootout won't hammer the CPU all that much.
Keep in mind that every game is different. A wide variety of games will need to be profiled. Don't run the most intense game available, profile it, and think that applies to the whole system. Though if it's a game that's played daily, then it can have more weight to your decision making.
Well crap, I don't think I should upgrade because the video card will be bottlenecked.
The question to consider is how much of a boost will the system get anyway? What about your upgrade plans? If you plan on building a system later but you can get something now, why not get the upgrade and reap the benefits? Ultimately it's up to you, but don't think just because your system has a high chance of bottlenecking a future upgrade doesn't mean you shouldn't upgrade at all. Set some (realistic) requirements like the system must perform at least 45FPS at 1080p in most games on their highest quality preset (not including AA). Don't sweat it if the system still meets or exceeds them. If your system is getting 40FPS on a game and an upgrade can do 80FPS, but on your system it can only do 70FPS, but the next card down can pull about 60FPS and isn't all that much cheaper, why not still consider it? (I don't know of any real life situation like this, but just some food for thought)
I don't want to bother with profiling my games, just give me some pointers
I can't seem to find a lot of websites that aggregate their CPU benchmarks on games with actual figures, are fairly recent, and have a wide sample size. What I did find was Anandtech's CPU benchmarks (http://www.anandtech.com/bench/CPU/1357) and a slew of articles from TechSpot (http://www.techspot.com/features/gaming/gaming-benchmarks/) along with some various other articles or sources that had at least a large gap in performance. (such as this video putting a GTX 1080 in an i5-750 machine) So based all this, here's my conclusion:
In most cases, you can go down to an Intel i3-4000 or AMD FX-8000 series and still be at least 80% performance the card can achieve (at least with a GTX 980/980Ti)
Even in this situation, you'll still likely achieve 60FPS or above at 1080p on maximum or at least high (it's also getting harder as of late to distinguish between the two, so I wouldn't sweat it if you have to tone down the quality a bit).
You're basically in the danger zone if your processor at the time of release was about $100-$120 or less or is less than about 1900 or so on this chart . Though this only applies to DX11 titles. There's not enough data to make a conclusion on DX12/Vulkan, but so far the trend is more cores is better if the game uses async compute (not all games will!)
The tip I want to leave with you is this: bottlenecks aren't the end of the world. Determine how much you're willing to put up with.