Cupla added thoughts ....
A. One thing to not going forward is that CLC type water cooling loses its efefctiveness over time. This is inevitable. Unlike OLC type AIOs and cusrtom loops, the CLC folks violate the 1st rule of water cooling ... "Never mix metals in a cooling loop" . This sets up a galvanic corrosion cell due to the wide difference in galvanic potential between the 2 metsls"
Nickel = 0.30
Copper = 0.35
Brass / Bronze = 0.40
Aluminum = 0.75 - 0.95
With "real water cooling", we are typically dealing with small differences in galvanic potential... typically at worst, 0.05 - 0.10. With aluminum, we are looking at 0.45 - 0.65. In a galvanic cell, the "less noble" metal "aluminum" will shed electrons which cause deposits to from on the "more noble" metals to which those electrons are drawn. Anyone who has ever owned a boat is familiar with this process as that's why you attach sacrificial zinc anodes
http://www.boatus.com/boattech/articles/marine-corrosion.asp
The problem created is two fold. 1) your radiator, especially at soldered joints will start to corrode away ... this will likely not be readily observable right away and is usually hidden by paint... 2) is that the biuld up or deposits in the blocks reduces air flow and insulated the heat movement from the copper baseplate to the coolant. If you monitor temps over time, this will usually be the first indication that things are changing. You can the effects of 4 years of usage here .... warning nerds and geeks my experience physical discomfort when viewing these photos:
https://martinsliquidlab.wordpress.com/2012/01/24/corrosion-explored/
The issue is significantly exacerbated on CLC type sealed systems as being "no maintenance", you can not augment the initial corrosion inhibutors which have a useful life of only 18 - 24 months. In a custom loop or open CLC which allows the user to add additional inhibitors, you can perform "recommended maintenance" just as you do on your automobile's cooling system/. Such inhibitors do have anegative effect on thermal efficiiency and while we are generally willing to give up a few degrees on our cars, PC enthusiasts are generally less willing so smaller dosages are used which is why the recommended interval is 18 - 24 months.
So in the end, if using a CLC cooler like the Corsair's (basically anything OEM'd by Coolit and Asetec), Id recommend collecting data under specific loadings (i.e. RoG Real Bench), recording ambient as well as all pertient tpay more attention. If you want to avoid this situation on next build and don't want to do a custom loop, I'd recommend the AIOs from Swiftech (all copper) or EK (all copper or all aluminum). In addition to no dissimilar metals, you can add additional blocks (MoBo, GFX card, etc) additional radiators, reservoirs and anything else. The EK is real simple if ordered prefilled, tho I prefer to install myself when doing user builds.
Another little trick is to place a valve or at least an elbow w/ cap on the radiator drain port. 1) It allows you to easily drain and replace coolant... or 2) add more corrosion inhibitor at some point. Simply turn off the system, take the cap off the elbow, add a few drops of inhibitor and reapply the cap.
B. The CLC's typically have very low volume pumps ... the H100i was measured at 0.11 gpm... whereas custom loops and the OLCs mentioned above are 10 times that. As such, when you get any buildup, the pump does not have the ability to 'push' past it. You should also watch pump rpm and see if that's changing over time.
C. Never read temps in BIOS because you are not in fact "at idle" ... Idle, by definition, is when your system is at no load and Windows power saving measures have kicked in. If you load HWiNFO at startup and look at CPU speed, it will often remain at full OC multilier ... it won't drop down to 800 MHz till after about 5 minutes
hahaha joke bro... Oky 
