Jonathan Lemmens

I think its the NH-C14(S) though, its a 140mm cooler, slightly larger than the NH-L12

1. Yes static pressure fans are better in front of a radiator or a thick fan filter.
 
2. There will not really be a 'battle' as you say (unless you have no intake at all). The reason why the card trick works is because it keeps the pressure in the case slightly positive, which helps with dust management and forces more air trough the rad. But, this is only valid if you have a neutral/positive pressure in your case to begin with, i.e.: equal or more intake fans than exhaust fans.

I would not consider Cinebench a real load.  While it's indeed tough on the CPU, it's for a short duration.
 
If you have any concerns with your PC actually being stable, I'd suggest running a more aggressive stress test application for at least and hour or two. 
 
Don't worry too much about 3.8 or 3.9 GHz.  The difference in performance isn't worth being fixated on 3.9.  If you have to drop to 3.8 to obtain stability or safer voltages, you're only losing a theoretical max of 2.6% in performance.  In reality, probably less than that.
 
Good luck man.

I'm not saying 4Ghz is out of reach, there are a lot more factors involved in getting a stable OC. Though it is clear you don't have the greatest R7 1700, most BIOSes have a few more settings you can tweak to increase stability, maybe it could even be the RAM that causes the crashes, though I doubt it. 
 
See if you can find an OC guide for your specific mobo that explains all features and what each setting does, if you're not very experienced with this.
 
Edit: which mobo do you have?

Yeah i'll have a look for these VRM settings. Thanks for your help.

RAM is more often than not an issue with limiting performance, unless it's a capacity issue. But it really depends on the game.

If power consumption means nothing to you, Freesync is worth it imo. On the other hand Vega will not be as good as the 1080 Ti.

Bottleneck depends 100% on following points:
 
- Your Software
- Your Settings
- Your expectations (how much fps do you "want" to get)
 
 
You "only" have an i5 CPU. For 60 Hz ( = 60 fps target) that's fine. But in Mind fo 144 hz, you might struggle to get 144 fps, even if your GPU would be fast enough for it. A mere i5 won't.
If you reduce Graphic settings to achieve 120+ fps in all your games, you will probably sit in a CPU limit in many games.
 
BUT: Even so, ask yourself 1 Question: Are those fps, your CPU is able to deliver, enough for YOU?
If YES: No problem. Who cares if there's a small Bottleneck here, or there. YOU have enough fps, everything is fine. Enjoy your game.
If NO: You will have to upgrade your CPU. Bios update your Motherboard, get an i7 7700k, and sell your i5 6600k. Won't be that much of a Price then. And you will be able to push much higher fps in those high regions of 120+ fps.
 
 
Sorry, Edit:
 
You do have a monitor, that supports Freesync. Forget Nvidia.
You are underestimating VRR (Variable Refresh Rate) damn hard.
 
Wait for Vega, get Vega. Be happy.
 
Depending on Games/engines, and how much YOU are sensible to Tearing / stuttering.. you WILL notice Freesync (or G-Sync) aLOT.
In some games, having G-Sync on vs. Off is for my Eyes a difference like Day and Night, despite me having a 165 Hz Monitor.
 
You have Freesync ---> Use it. You will NOT regret it.
 
Having like 80-90 fps with Freesync enabled feels MUCH smoother, than having unsynchronized 100-110 fps with Screentearing/stuttering 
 
That means, even if a GTX 1080 ti might deliver 10-20% more fps on paper, a Vega 64 might give you subjective smoother gameplay, thanks to Freesync.

Aida64 for stressing CPU
HWMonitor/RealTemp for monitoring temperatures
CPU-Z for various system specs
GPU-Z for GPU specs and stats
3DMark for system/gaming benchmarks
Cinebench for CPU benchmarks
Unigine Valley and Heaven for gaming type benchmarks
CrystalDiskInfo and ATTO for storage benchmarks

When you delid you should always put the heatspreader back on the die with preferably liquid metal in between. That way there is no height difference.
 
The water temperature in a cooling loop is pretty uniform, with negligible difference in temp between incoming and outgoing water in the radiator. So placement shouldn't matter much.
AIO manufacturers usually specify a maximum fluid temperature, I think its around 60C. 
 
Edit: but you don't have an AIO I guess? The only component in a custom loop that is limited in its operation by temperature would be the pump, check the specs on that for more info.

It doesn't rely on core count or hyperthreading.  It simply scales almost perfectly with them as it should.  
 
Cinebench is a great benchmark for determining CPU performance under perfect scaling conditions.  

Thank you so much for the help.

When you delid you should always put the heatspreader back on the die with preferably liquid metal in between. That way there is no height difference.
 
The water temperature in a cooling loop is pretty uniform, with negligible difference in temp between incoming and outgoing water in the radiator. So placement shouldn't matter much.
AIO manufacturers usually specify a maximum fluid temperature, I think its around 60C. 
 
Edit: but you don't have an AIO I guess? The only component in a custom loop that is limited in its operation by temperature would be the pump, check the specs on that for more info.

What? My expectation of stability is rather simple. I just take the most convoluted path to get there
 
 

Ahh excellent, thanks a lot. I've learnt a lot in just 6 hours. Considering i had no bloody idea what i was doing at the start.

This is not necessarily true, this chip might still reach 4Ghz at 1.5V for example, but no sane person would do that on aircooling.
 
You're right that lower temps doesn't mean more Ghz directly, but lower temps does allow the CPU to be stable at lower voltages, which can sometimes give a little more OC headroom.

The CCleaner registry cleaner is actually pretty good. It's great for cleaning up junk left behind after uninstalling products that don't properly clean up after themselves.

To add to anything mentioned above I would add your favorite games, especially if they are demanding ones. 
 
Also: Benchmarks are used to determine PC performance. 'Bench marks' are left by people who didn't wipe their ass properly.

ya ok. i don't need it i already have a skylake 6600k and an i7 920 1366 (witch would probably beat that server for gaming even thow its only;y 6gb ram) it be just for messing with or sell it as a server / buget gaming pc. (as of right now price dropped to $150 so i might be able to get it cheaper haha.
 
the problem right now is everyone is listing there old pc's for too much. like my i7920 1366 pc is worth $250-$300 even thow its like 9 years old. come to think of it i could probly resell my ssd i got in December for $135 for a 480gb i could probably resell that for more then i paid because the price dubbed.ahah
 
so i was watching a youtube video on guy making an mindcraft server and he said it would work with 1gb of ram 2gb ram for moded. i guess for servers you alacakte some ram to a server. the more ram you have the more servers you can run is how i see it i guess. so in terms that seem like a good server haha.anyway thx for your time

Thanks dude very much

Of course it isn't, and it won't be: one is a dual channel CPU, the other is a quad channel, quad socket server. 
That's the whole point, and that's why people saying "better Ryzen!" are wrong. 
We're not debating who has the better IPC. We're just discussing the fun of building this very specialized machine. (that's also why it has to be cheap). 
Isn't that what enthusiast do, playing with all sorts of computer stuff and learn, especially if they haven't tried before? 
Or has marketing convinced us that "enthusiast" means buying the most expensive toys even if we have no clue about what we're doing?
 

The GPU does not require 300W, this is an estimated minimum wattage required for a complete system with a 1050Ti in it.
 
The HP elite 8000 comes with either a 240W or 320W PSU, so not enough or just barely. If you have the 320W version you could use a Molex/SATA to 6-pin adapter but like MrJoosh said it isn't really recommended, especially if you have a PSU so close to minimum specs.

Im aware of what that card requires (75 Watts) and i've also tried this method with my Quadro FX 3800 which requires 110 Watts and it totally did'nt work thats why i want an alternate tower which can just support a genuine connector.  

No need to yell.
 
I'm quite confused by your terminology and your questions aren't very clear either. Also in your initial post you said "there is no budget" so, good luck finding a new PC without money.
 
I'm happy to help but could you please rephrase the question and perhaps provide more info on the hardware you have (maybe a photo), your budget and what you want the PC to do?

So you're asking what case is best?
 
You mentioned earlier that workstation cases often don't accommodate ATX PSU's, with that comes the issue that most proprietary mobo's often don't have standardised mounting holes and often non-standard power connectors as well. On top of that the mobo itself is maybe a weird shape which may not fit an ATX case regardless of mounting holes. So you'll probably encounter more problems to build this system in a different case than 'modding' a regular PSU in this case.
 
If you at least have standard ATX power connectors on the motherboard your best bet is to buy a cheap 450W ATX PSU and, given the budget, just mod the case to make it fit.