I edited and numbered your post so I could address appropriately without too much clutter.
1. Ok... so now we doing a gaming build only... so that means LGA 2066 isn't where we wanna go. 28 threads isn't going to do anything for you, You won't see more than 4 in use 97% of the time.
2. 1000D won't take 140mm wide rads in front
3. Two separate loops is not something Iid recommend recommended ... for the CPU with it's small surface area... you'll want to move coolant thru there at 1.25 gpm or better .... on the GFX cards, you are going to have HUGE full cover water blocks with large surface areas so you will want to use half that flow .... so the easy way is run a single loop up to the GFX cards, then split the loop into 2 separate parallel flow stream and recombine right after the GFX card blocks.... allows you to increase flow as needed without generating excess backpressure by only having half the flow thru GPUs )or 1/4 the pressure) .
Keep this i mind... Your two GFX cards are 320 watts each ... total 620 watts... think about that ... 1 loop for your 200 watt CPU and 1 loop for your 700 watt GPUs. Leaves one overdesigned and one underdesigned.
4. Again lets look at 820 watts of heat load
https://www.overclock.net/forum/61-water-cooling/1457426-radiator-size-estimator.html
Open the Aquacool speadsheet
Read the instructions... recognize that all surfaces on the components are giving off heat ... all MoBo surfaces, block surfaces, tubing , radiator shrouds, etc will account for a good part of the heat load... also note that the CPU is never at 100% when the GFX cards are.... so combined this will account for 40% of you heat load, leaving 60% for the radiator. So...
60% x 820 = 492 watts or 246 per rad.... now look at the spreadsheet ....
With just fans in pull, a 45mm thick 480 will provide 245 watts of cooling at just 1250 rpm ... since ya like to "Go big", use a 60mm thick one for 251 watts ... 303 if you go push pull.
As per last post .... make sure ya have 1.5 intakes or each exhaust....but each 140mm = 1.33 120's
Front .... I would start with 8 fans in pull (4 per rad) ... as you can see in the spreadsheet, you don't need push pull. And you certainly don't need a top rad. Can always add the Push later if ya think ya need it ...Here's how I'd run my loop....
Two Radiators =>Separate Discharges => Inlet Y Fitting => Dual Pump => MoBo Mono Block => CPU => Y Fitting => Separate Feeds to each GPU => Y Fitting => Separate feeds to each Rad
This way 1) You have the symmetry you like.... 2) you reduce backpressure and flow thru the rads and GFX card WBs allowing greater cooling.
5 Not recommended. Again, rad fans ALWAYS blow in, no exceptions ... ever. Put 8 fans on rad blowing out and 8 blowing in ? Well since the ones with inlet filters in front will have flow reduced by the filters, you already have negative pressure.... now put (3) 140s on top and (2) in rear blowing out and you have very negative poressure. Your air flow thru the box will be in excess that the entire volume of the case will be pushed out 2 or more times per second. Those giant wide open grilles on back of case are part of your cooling system. And do you want to cool yor stuff with cool ambient air or pre-heated inside case air ? With coolant at 33C, you want say 23 ambient air cooling ya stuff ? ... or ya want interior 28C air cooling ya stuff which means
only 50% of the cooling (33-23 vs 33-28) ? If ya think that is bad for interior components, go thru your parts manuals and find one component which component is going to be bothered by 28C ?
6. Not sure how ya came to conclusion that I did not allocated only one radiators worth of fans,..... each hub handles 11 fans.
Radiator 1 Pull Fans (4) + Radiator 2 Pull Fans (4) = 8 fans out of 11 ports filled (I would install this at time of build)
Radiator 2 Push Fans (4) + Radiator 2 Push Fans (4) = 8 fans out of 11 ports filled (I'd save this if later ya think you need it.)
That way you can put a lil more rpm on the pull fans if ya want to .... or even have 1 set of fans kick in at light loads and not add the push fans until temps reach a certain point.
7. Unfortunately, forum advice from yesteryear continues to be parroted when the reasoning behind those old recommendations no longer applies. Back in the day, rads were 30 fpi and the closeness of the fins was harder to push air between. Today rads are typically 8 - 14 fpi and there is no need for medium (M) and high (H) static pressure fans on today's rads. An AIO with high fpi and cheap aluminum rad will need extreme speed fans to make up for the aforementioned shortcomings. A properly designed all copper loop, designed for Delta T of 10C (all those spreadsheets are based upon Delta T of 10C) ... AIOs are like 20C or more. As you can see in the spreadsheet, 2 x 480 with just pull fans is more than enough for your load.
8. No, absolutely not... again, no exceptions... rad fans ALWAYS blow in unless you have a poorly designed case. Look at Corsair AIO instructions.
https://www.corsair.com/corsairmedia/sys_master/productcontent/49-000175_rev_AB_H100i_QSG_web.pdf
read page 3, Window 1
"For
best cooling performance , we recommend
mounting the fans as air intakes into your PC case". You have to forget what you think you learned in 8th grade earth science. Yes hot air rises but not with a fan over it blowing the other way .... If ya worried about cooling GFX card and CPU, why in the world would you want to reduce the performance of your cooling system by preheatting the air before it goes thru the rads ?
10. Two issues here:
a) All things being equal (design, rpm, etc) , a 140mm fan blows 33% more air than a 120 mm.
b. This is kinda like a Dad telling his son that Santa Claus isn't real. I hate to be "that guy" but the specs you read on the manufacturer web pages and on the Box are blatant lies. Lets look at WC guru martin on the subject
https://martinsliquidlab.wordpress....w-specs-are-poor-measures-of-fan-performance/
Here we see two fans:
Cougar Vortex PWM
Max Airflow = 70.5CFM
Static Pressure = 2.2 mm H20
Gentle Typhoon AP-15
Max Airflow = 58CFM
Static Pressure = .08inwg = 2.03mm H20
Which one's better ... isn't 70.5 bigger than 58 and 2.2 bigger than 2.03. That would be true if those numbers were real. But, what Cougar is saying is ..
At 0.00 Static pressure, our fan pushes 70 cfm, at 2.2mm static pressure, our fan pushes 0 cfm ... both conditions will never exist.
What we see once installed is ...
The Cougar fan produces 34 cfm at 0.040 in of static pressure (42% of the advertised flow)
The GT fan produces 38 cfm at 0.047 in of static pressure (67% of the advertised flow)
11. Internals
MoBo / CPU 1x Asus ROG Rampage VI Extreme Intel i9 9940x 14 Core, 28 Thread - Again LGA 2066 has no business in a gaming only box.... most reviewers don't even do gaming tests. The question is not whether it is good or bad in gaming but whether it's better than something that costs less than 1/2 or 1/3 the $1400 cost of that CPU ..... If ya want the very best, most cores , most expensive gaming CPU , get the $570 9900k
hhttps://tpucdn.com/reviews/Intel/Core_i7_9700K/images/relative-performance-games-2560-1440.png
Personally, Id spend $150 less and get the 9700k for a gaming box.... likely going down after holidays to mid 300s
As for Mobos, you could "go big" with the MSI Z390 MEG Godlike for like $580 ... but I don't see what I'm getting for $300 over the MSI MEG Z390 ACE to make me part with that amount of money ... 2nd ALC 1220 ? ... 2nd Killer E2500 Gigabit LAN controller ?
4x Corsair - Vengeance RGB Pro 8 GB DDR4-3200 Memory (32GB Total)
a) I have yet to see a game improve with more than 16GB
b) 4 sticks will likely reduce your OC as compared to 2 due toi the extra load on memory controller.
1x Samsung 970 Evo 2TB M.2' - If ya got the cash, why not ?
2x Western Digital - Gold 10 TB 3.5" 7200RPM - I have 30 years of business records and AutoCAD drawings, all my business and personal tax records, photos, games and backups of 5 networked PCs on a single on 2 TB. We have not purchased a HD in 8 years. But if you need the space, get what ya need. If it was me, id use an NAS in RAID 5/10 .... Be aware though:
a) Large capacity drives have a much higher failure rate.
b) SSHDs are 50% faster in gaming than WD Blacks
c) Seagate SSHDs (0.44%) have hair better failure rate then WD Blacks (0.45%)
d) Both have 5 year warrantees
e) Seagate tops out a 2 TB, WD has a 4 TB
12. Some games benefit from lower timings, some games benefit from more speed.... most times however, they don't test with twin GFX cards where often RAM can make a difference .... with just the 1 cards it is bottlenecked by GFX. However, with your 1440p monitor, that's not going to be the case. The sweet spot is DDR4-3000 CAS 15 / DDR4-3200 CAS 16 ... after that proces rise significantly, ... if ya can find better for a good price, grab it.
13. If you use benchmarks as a yardstick, then the more expensive SSD can be justified. If you use application benchmarks , using scripts, faster SSDs can be justified. If you use productivity as the yardstick, SSDs have no real advantage outside specialized apps like animation, video editing and rendering. Boot time of an SSD is 15.6 seconds versus 16.5 for an SSHD. Users are not in a position to take advantage of that. If a game takes even 10 seconds longer to load, by the time I "do what i gotta do" to switch ME from work mode to gaming mode, either way, puter is waiting for me to unplug headset from charging cable, pull out dongle from storage cubby, close cubby, insert dongle, load discord, load web sites ... and that doesn't even include bio break, snackies and stuff. So while there may be some sense of personal satisfaction, no one every typed a extra legal brief, reached a further waypoint in game whatever because they had a faster SSD.... best advice.... buy the best ya can till what ya paying causes discomfort.
14. RAID 0 hard drives are a complete waste of time, money and effort. Every 3 years we build a box with RAID 0 and RAID 1 arrays as a test bed. Every time we have broken the arrays as they returned 0 benefit by every measure imaginable. Even for SSDs
https://www.tomshardware.com/reviews/ssd-raid-benchmark,3485-13.html
https://hardforum.com/threads/raid-...desktop-application-game-performance.1001325/
https://tweakers.net/reviews/515/9/raid-0-hype-or-blessing-conclusions.html
Most accurate statement I ever read on RAID 0 was that those who pursued this route were ""buying into the hype and nothing more".
Aside from them charging more for the 99xx series (just gauging results from Scan.co.uk) I can't see the importance of buying one. The only reason it's slightly faster also, is down to the slight increase in clock speeds.. If you're overclocking, it appears to be a waste of money to me.. Plus that and the de-lidding issue, personally it's hard for me to recommend..
