Hulkie asked me to stop by and add a pic .....
Waiting till I get the 3rd Phanteks PCB befopre installing final fans and doing the sleeving.
-Phanteks Enthoo Primo Ultimate Chassis - puuurfecr case for M6F and WC
Seasonic X-1250 - Edged out the AX1200i and was $100 cheaper at the time
-Asus Maximus VI Formula - Cuz it's gorgeous
-Intel 4770k - @ 46 multiplier / 46 cache ratio on BIOS 0804 ....45/42 on 1102
-(2 x 8GB) Muskin CAS 10 DDR3-2400 = Perfect color match and best timings on market
-2 x SAMSUNG 840 Pro Series MZ-7PD256BW 2.5" 256GB - Started in RAID 0 but was disappointed, now individual drives.
-2 x Seagate ST2000DX001 2TB MLC/8GB 7200 RPM 64MB Cache Solid State Hybrid Drive -Was RAID 1 post build, now separate drives
-2 x Asus GTX 780 DCII w/ 25% OC on core / 20% on memory
-Mushkin Ventura Pro 64GB Flash Drive
-ASUS VG248QE Black 23" 144 Hz Monitor - waiting for G-Sync Module
-Logitech G19s 920-004985
-ASUS Republic of Gamers GX950 Laser Mouse
-10 x PH-F140SP_BK (5 LED variety) w/ 5 more coming as soon as I can get my hands on a 3rd Phanteks PCB Fan controller
-Swiftech 35x2 Pump
-Alphacool 280 + 420
-EK Supremacy CPU Block / EK Full Cover "Clean" GFX WBs
-EK Res-3 250
Will try and address many of the questions raised in the thread to date....
1. All BIOSs after 0804 have had reported issues with formerly stable OCs going unstable. Also true on Hero.
2. Once I set puter up on LAN, WiFi stopped working....got it back at cost of LAN..... disabled WiFi as not needed after puter reaches its home.... only used WiFi really during build stage (no service at my workbench aka dining room table).
3. Only CPU and CPU_OPT headers are PWM. CHA fans are VVC (variable voltage control) as are OPT Fans.... FanXpert2 controls CPU and CHA fans.... you can not control OPT fans with the thermal sensors via FanXpert2 but I believe you can to a very limited extent via the BIOS.
4. All CHA and OPT fan headers are rated at 1 amp. I would not put more than 0.50 - 0.67 amps on them with 1800 rpm and up fans so as to allow for inrush current which is proportional to starting torque which will be higher the faster the fan spins.....with slower fans (i.e. 1200) inrush isn't that bad so I'd go with 0.67 - 0.80 amps. If using speed control, logic dictates that ya can get more as that substantially limits inrush current. FanXpert 2 works great BTW.... Once I get the 3rd PCB Fan Control from Phanteks ....
-CPU and CPU-OPT control the two pumps
-CHA_1 controls the (5) case fans
-CHA_2 will control the (6) Rad Fans on the 420
-CHA_3 will control the (4) Rad Fans on the 280
Until the 3rd PCB arrives, have the (5) push fans on CHA_2
Fan speeds range from 325 - 850 rpm and remain dead silent unless manually set at full speed.
CPU Core Temps top out at 74C at 46/46 under 0804 BIOS under RoG Real Bench, lower under IBT
GPU Core Temps top out at 44C under Furmark w/ fan control, 39C if set at full speed of 1200 rpm.
Pump speeds range from 1400 - 3000 (max is 4500)
5. Speedfan does not work on the M6F
6. The key to using FanXpert2 is to have run the configuration program and have it test / profile all installed fans (well at least those connected to CPU / CHA headers.
7. Not only can't ya do tri SLI .... but even a sound card in that 3rd slot kills x8 x8
8. The Asus MoBo Block is not nickel plated but anodized aluminum to which a coating has been applied. Do NOT confuse anodized aluminum with aluminum. Having over 25 years experience with Anodized aluminum on severely corrosive environments, the mild exposure here with DW or engineered coolants presents no concern whatsoever.
9. If you are exceeding 1 amp on any particular header, you can use a PWM splitter which come ion two flavors:
a) The Swiftech and ModMyToys PCBs take a PWM heed from the feader and a separate 12v Molex feed from the PSU to control up to 8 PWM Fans. Both are available from FCPU
b) The Phanteks PCBs take this a step further by taking a PWM feed from the header and a separate 12v Molex feed from the PSU to control up to 11 VVC Fans. Right bnow only comes with Phanteks cases but will be available end of this month as a separate accessory.
10. The Dominator 2400 RAM is nice but the Mushkin Redline Redbacks have the same timings, are cheaper and a better aesthetic match. Mushkin is till using Hynix modules but I know Corsair switched to another brand on the Vengeance Pro, not sure on the Dominators tho. I have oft found 2133 for cheaper than 1866 / 1600 and the 2400 barely above the 2133..... 2666 tho prices are a bit higher so wouldn't go there. In my experience, faster RAM has only single (2-5%) digit impacts on average fps but minimum fps can go up as much as 10% on some games..... of course if ya use rig for other purposes (i.e. video editing, CAD, etc) faster RAM's cost is easily justified.
Now when a line of RAM 1st comes out there is a significant cost premium between speeds and a year ago I woulda said not worth the investment. As stuff comes off the line, yields on the high speed stuff are low and therefore the higher speed stuff is rare and has a significant cost premium. What doesn't pass 2400 will be tested at 2133 and if it fails, will be tested and sold as 1866 when in passes. Which is why ya see so many reviewers saying it's not worth the cost when reviews are done soon after release..... and these reviews get parroted in forum messages long after what they said was relevant. However, as the line matures, yields improve and the hi speed stuff is common and therefore cheaper.as often as not I see 2133 at the same or even cheaper than 1600 or 1866. Now does it make any sense to listen recommendations in an old review that recommend "get 1600 cause anything faster provides only a small benefit" ? Not when 1866 and 2133 are same price or close to it. The last time I bought 16GB Mushkin Redline Ridgebacks (2400 @ 10-12-12-28), they were $15 more than the 2133's which were $5 more than the 1866s and same price as the 1600s.
Now the 1866s are oft the cheapest so a no brainer over the 1600s. Last time I priced them the 2133's were same as or $5 more than 1866s. To my mind, if faster RAM is gonna cost ya $20 or basically 1% increase in system costs .... for that even a 2% increase in performance in the most mundane of operations, is still a 2:1 ROI. However, sacrificing CAS for speed can be a bad idea... and of late getting lower CAS oft has a higher cost than higher speed.
Time = CAS * 1000/DDR Speed, so....
CAS 7 / 1600 = 4.38 ns
CAS 8 /1600 = 5.00 ns
CAS 9 /1600 = 5.63 ns
CAS 8 / 1866 = 4.29 ns
CAS 9 / 2133 = 4.22 ns
CAS 10 / 2400 = 4.17 ns
CAS 11 /2400 = 4.58 ns
In short, whether or not it's worth buying faster RAM depends on both DDR Speed and CAS but most importantly, the price structure at the time you buying. A year ago.... I was buying 1600 cause I couldn't justify the increase in performance for the increase in price..... now I only buy 2133 / 2400 because the price difference id oft just $0 - $20
11. Sound is great ....equivalent to the $89 Xonar. All drivers and utilities (i.e sonic radar) come on the MoBo CD as well as can be downloaded from Asus website.
12. AIDA64 / Prime96 testing is NOT recommended an any OC setting other than manual.....RoG Real Bench is ideal for testing under Adaptive or Offset. You will note that testing under RoG Real Bench Image Editing portion that you will see a performance hit when cache ratio is more than 3 below CPU multiplier. At 45/45 .... down to 45/42 I saw minimal impact but testing at 45/39 I dropped about 3 %