I do have a fan wall with 2 Phanteks T30s on it so there is a good amount of air moving over the motherboard, but the PSU is making the CPU cooler's job harder:
I was planning on fan swapping the A47's fan to a 9RA0612P4J0011 If I got one, which should be in the same ballpark performance-wise, but vastly more quiet.
Oddly enough, I was able to buy an SNK-P0094AP4; I figured it'd be easier to fit and easier to quieten down the fan on this than it might be to retrofit a minimum-700rpm to the NH-D12L. Information on the SNK-P0094AP4 isn't easy to come by but in my notes I think it has a max height of 149mm so I don't think it'll fit in a 3U rackmount.
I'm probably budgeting for 100W board draw for when this thing's going at full pelt (lies, damned lies, and TDP figures) since I think the TDP will only be remotely realistic if boost is disabled. The dynatron is guaranteed to be a proper turbine though
Yeah, I think that might not be a good setup for cooling. The CPU fan isn’t the issue here. Those wall fans max out at 3000rpm, are pretty blocked from exhausting to the rear, and lastly from what I can gather from the picture, there’s almost no rear ventilation. Your fans are probably mostly just circulating hot air around in your case.
I think you might have model #s slightly off. That cooler ships on their 2U server: AS -2015A-TR | 2U | A+ Server | Products | Supermicro
There is a rear exhaust that lines up with the heatsink, the amount of air coming out of it is about evenly split with the amount of air coming out the PCIe slots based on feeling the airflow with my hand... perhaps I can block the open area of the PCIe slots to force more air through the exhaust next to the CPU cooler.
I'm running a 7900 on my H13SAE-MF which is also a 65 watt CPU and hwinfo is reporting that it regularly hits 95 watts of power consumption.
I chose the Supermicro board because it had the best PCIe lane layout. That being said Gigabyte isn't on my buy list any more because of the amount of hardware bugs in their boards that require board revisions to fix. I still consider Asrock Rack boards but their BIOS support can be a little bit lacking and availability in isn't as good as Supermicro.
Received my 4565P from Newegg this afternoon . . .
Three weeks ago, the H13SAE-MF was In Stock on teh Egg. I waited to Purchase till the next morning, and OOS!!! About two weeks ago, there were ~11 H13SAE-MF available on teh Bay for ~USD520 . . . that dried-up the a day to the week after I decided to pull-the-trigger on just one.
Learning to appreciate delayed gratification is one of life's most important and least-celebrated adventures I feelWe're probably competing with the interests of approx. 15,528 other purchasers, so try to remain gentle with your Self
As we'll soon be suffering from August, that'll mean I'm on holiday anyway.
Received my H13SAE-MF [REV: 2.00] board, and E610-XT2 card, todayLearning to appreciate delayed gratification is one of life's most important and least-celebrated adventures I feel
Just keen to get some time to have a fiddle with it. My current 3700X has done me well the last five (nearly six?) years, and will probably still be completely usable for another five, just want to see how much the bang-for-buck swingometer has moved in the interim.
Are you running x16 cards in both of the full size slots trying to get x8x8? I remember reading that the stupid mux chip can refuse to do x8x8 if the primary card is x16 and someone had to use tape over the latter half of their pins.
I have the E610 in the top x4 slot (which works well afaict), and one A4000 4.0 x16 in one 5.0 x16 (merely tested both).