they arent visible with the side panel on, or they'd be there
I'm just bustin yer chops man.
It looks great.
I've been running this 4.2ghz, 1.3v overclock for a while now but lately I've been thinking of taking it down a notch. Sick of my fans kicking up for everything. The voltage jump needed to hit 4.2 vs 4.1 is rather high, and so is the temperature gain. Under the meanest, heaviest stress testing that 4.2ghz overclock will push things up to ~75C. Which isn't terrible for me, but it's the limit of what my cooling setup can handle. Not that I ever hit that - the closest I get is when my heavily compressed backups run. For gaming and audio stuff I may see 60C at most.
The really annoying thing about this board is that fans are actually tied to socket temp instead of core temp, so I have a much narrower region to set my curves to... ...between 28C and just over 60C (SOCKET temp, not CPU.) Seems like a minor thing, but it really is a huge PITA. For whatever reason, and I don't know if it's this motherboard, or just the combination of this board and the 2600, but in HWiNFO64, the sensor marked CPU is socket temp, and that's what the fans follow, whether via BIOS or software. It's so beyond stupid, I can't understand...
And since it's not responding to core temps, once the socket heats up they just run and run, even if the CPU has dropped from 80C to 60C. So to keep up with that 4.2 OC, I have to set them to basically max when the socket hits just over 60C. But the CPU could be significantly hotter or cooler than that at any point. Fans won't know that. The socket doesn't really get hotter than 65C no matter what. Basically makes it so I have to make the fans ramp up towards max speed way before I need them to just to have the air when I really need it. But what this means is that when I'm gaming and hitting 55C, the fans are running at like 75% or 80%. And that 80% isn't doing more than 60% would do for it. And then when I stop the game, the fans will continue to run hard for a solid minute after, while the socket cools, even though the CPU has long since dropped to its 30C idle.
The flipside is that say my CPU jumps to 75C quickly. It will take as long as 10 seconds before the fans rise to meet the needs of that, because the socket is still warming up. It's possible it wouldn't hit that temperature to begin with if the damned fans had just kicked up when they were supposed to
I could set up a bunch of different fan curves, but I like stuff to work one way and have that one way work for everything. Once it's setup I shouldn't need to mess with it to keep it working like I want it to.
Or I could set it to have the fans "cap out" at whatever speed I want for my usual usage. Set the fan curves to max at like 75C and then put a dot at 60C running whatever percentage I want. Since the socket doesn't go much over 60 the fans wouldn't speed up past that. And yet as long as they're spinning fast enough at that point, they'll keep it cool and reasonably quiet. But somehow I don't like that. I just think, well, what if I ever do run something more demanding? Then I'm in for a thermal shutdown situation.
So I said screw it, swapped MX-4 for Kryonaut, and dropped the CPU down to 4.1ghz @ 1.225v. Biiig step down for minimal performance loss. Now IBT only pushes me up to 63 or 64C. Small FFT's plateau about 61C. And now I never have my fans maxing. I barely even hear them. Temperature still went down while playing Fallout 4, even bringing the fan curves down. The room also stays much, much cooler during long gaming sessions
Funny how that happens with core clock and power, though. 4.1 and below, temperature, current, and voltage go up in small increments. But take it up to 4.2 and suddenly the voltage has to go way up, power usage goes up by at least 20W, current draw increases by like 20%, and temperatures go up at least a full 10C. I'm assuming that sharp efficiency drop has something to do with infinity fabric. It's just... ...it really is an insane amount of heat to get 100mhz.
), comes when you install F22 UEFI update. Only thing i touched/enabled was Custom Pstates/Throttling P0 option, the rest @ Auto. Things to learn:
& no matter what i did, when @ Auto all well & good, 1.38v on vCore & temps are not fluctuating, once set to any voltage manually, 1.548v & stays there. Made me go in my head "WTF is this shit???!!!!!" Found workaround for this: unplugged the 24-pin power connector & plugged it into LCS's 24-pin adaptor thingy, removed the air bubbles (had to reconnect properly the LED RGB connector which involved temporally removing the res that stood in a way of graphics card screws & card itself, & it made the coolant move places, laws of physics abide 
), turned back on - 1.380v now & all snappy. Pic of RGB awesomeness:
