I don't understand.
At this point it sounds like you're just trying to bash me to win an argument that shouldn't exist. If anything, it should be a civilized discussion. But saying things like:
Wow. This is interesting, because throughout your post you acknowledge stuff I've said. Contradiction?
makes it look to me like you're trying to ridicule me like a kid in grade school. I'm not sure what your intentions are, but that's how I'm taking it.
I don't know how you're taking what I'm saying either, but I'm just trying to be helpful and give you and sevin a better understanding of how to choose power supplies, since you seem to have an understanding based solely on something you've read somewhere on a specific forum/guide. Why do you comment that whenever I gave non-related advice, it doesn't contradict what you said? Again, I'm not just trying to sit you down with everything I say, just help. If that's not what it comes off to you as, I'm sorry.
Anyway,
It's not what I've said. Regardless, there is an optimal wattage for a system. You measure the maximum power consumption, and that should be around 60% of the labeled wattage. That's where the efficiency curves have their peak (with slight variations ofc). Replace the 60% with 80% if you have something precalculated like sevin mentioned before, since those values tend to be higher.
Even though this is true to some extent, it's not something you should think about as hard as you are, since the estimation of your system's "maximum power consumption" is never accurate, unless you take days to collect data from many sources with the exact same hardware configurations as yours. Even then, the difference in efficiency at this level is so ridiculously insignificant that it doesn't matter, and you would've been better off with more wattage for future upgrades/more stability anyway. Or even better, a higher-quality power supply that wasn't available at your "optimal wattage". There is also so much variation between different models of power supplies that the estimation becomes even less accurate, especially if there is a big difference in how they allocate their power between rails.
If the "optimal" labaled wattage was ~450W and you could choose between a 450W power supply and a 550W power supply, and the 550W power supply had more stable power, less ripple, etc. it would be stupid to take the 450W one.
All correct, exactly my point. If this is "too extreme" for you, sevin is asking for help and I feel inclined to fulfill his needs as well as possible. Why not find a fitting PSU then? Because it's "too extreme"? There's obviously a point to it. No reason falsely call someone out just because you happen to disagree on the weight of the point.
Yes, it is too extreme. It's very misleading to act like choosing the "correct" wattage is very important, because it isn't. A good power supply objectively depends on so many more important factors, as both you and I have stated before. And again, a slightly "higher-than-optimal" wattage is usually not a significantly bad thing anyway, but rather a good thing. I think you may have been misled on this topic as well, are you experienced with building
modern computers?
It's not about picking the right manufacturer, it's based on the components which vary on a model-basis, therefore one should specifically pick out and look for a specific model and not for a manufacturer in general. I was in need of a new GPU just two weeks ago, and I had to go through a list of 970GTXs and find the cheapest one that has the smallest possibility (after all it cant be completely ruled out) of being affected. It turned out to be too expensive for my price range, so I settled for a specific R290X model.
I said manufacturer because most manufacturers only have one primary 970 model.
I'm sorry, but this is fanboy talk. Go onto any forum specialized in people handing out configurations and they'll smack you around for that.
Like a forum of AMD fanboys? But really, calling someone a fanboy for making valid purchasing decisions and leaning towards the brand that better suits their need based on their features is uncalled for.
AMD as well has complementary features for all of the above (most of which are actually open in contrary to physx), besides the point that these are all features that are either dead, barely supported or not needed. Some of the things you mentioned (mercury playback engine for example) utilize both CUDA and OpenAL and therefore are supported by AMD cards as well. (Plus AMD has Mantle *cough*)
So far dropping a list of buzzwords isn't much more than marketing.
ShadowPlay: GVR hasn't been around as long and isn't as tightly integrated (at least not yet).
PhysX: AMD has no equivalent support in games that support PhysX unless you want to hack it and run PhysX it in software (which is slow).
Mercury engine: I actually didn't know it supports OpenCL too, but I assumed it was Nvidia only since it only supports CUDA in Adobe CS6 (important).
CUDA: Some programs simply have better CUDA support and some have better OpenCL support. In this case, I think CUDA is more dominant.
3D Vision: I find it's way more solid and user-friendly than HD3D
G-Sync: Another situational example, but we don't have that much proof that Adaptive Sync is as good as G-Sync yet. It might matter a lot more in the future too when G-Sync monitors get less expensive.
Nvidia Shield: AMD has no equivalent.
Nvidia Grid: AMD has no equivalent.
And these were just examples. Of course AMD also has some nice features, but it's how well they suit a user's needs that matters.
Would have been 30 bucks but I'm sure there's cheaper z97s motherboards out there as well. Plus, which you should know, when he needs CPU upgrade in a few years, it's going to be on an entirely different plattform as well. I'm not sure how you perform your upgrades but you usually end up with longer cycles of cpu, motherboard and ram, and faster cycles of GPU switches in between.
True, but there is a chance that the next generation of CPUs will work on Z97 as well, and in whatever case, it doesn't change the fact that Z97 has more new features and is a better choice in general.
The boxed cooler itself is perfectly fine. The 775-Prescott days are over, they're actually decent since a few years ago. I wouldn't at all recommend some random person who seeks help in finding a basic configuration to go right into overclocking his new system. If(!) at some point later he needs the increase in CPU performance and if the boxed cooler shouldn't suffice anymore (mild overclocking should be okay, as long as he lays off the voltage since TDP raises linearly with clock and squared with voltage).
He is buying a boxed CPU anyway, which has the boxed cooler. He'll be fine. If someday he isn't, he can still purchase the cooler he would have purchased anyway.
The boxed cooler is perfectly fine for running a non-K, non-overclocked CPU, and I have used it in at least 3 different systems (2 Haswell, 1 Hawell refresh), but it just wouldn't be enough for any overclocking. I don't think you should do an insane overclock instantly either, but it isn't hard to do some decent overclocking or even let the motherboard do it automatically.
Even if you decide not to do it immediately, it's way easier to get the cooler first anyway so you don't have to go through the hassle of replacing the cooler afterwards and wiping off all the thermal paste. It will also make the system run a bit cooler and quieter from the start.
