I see your point as far as the benefits of having a single loop. Honestly, I hadn't really considered that because I was worried about dumping the CPU or GPU heat into each other. Especially since ideally you would want the top card to be getting the coolest water first, since that's going to be the card carrying the load when I'm not doing things that even take advantage of the other 3 cards, and even when running them at close to full load, really no matter what you do the primary card carries close to 2/3 of the load due to slot limitation in a quad SLI setup (correct me if I'm completely off base). However, doing a single loop with say RES>PUMP>CPU>MB Blocks>RAD1>GPU1>GPU2>GPU3>GPU4>RAD2>RAD3>RAD4>RES etc etc or even RES>PUMP>GPU1>GPU2>GPU3>GPU4>RAD1>RAD2>RAD3>CPU>MB Blocks>RAD4>RES etc etc (1st SEEMING better imo), I'd be looking at an extreme pump setup such as the Iwaki RD-30 (since running 2 seperate pumps in a single loop seems to be stressful on the pumps and not worth the cost for the performance you get) in order to get it to overcome all the restriction from the blocks.
Based on that, I'm not really sure (I am new to this and trying to understand) that in this case a single loop would be as much benefit as it would pain in the butt as far as trying to maintain enough flow. Now don't get me wrong, I've looked into more things and decided to go with a different CPU block to try to reduce as much restriction as possible. Cost isn't the issue, it's more about maintaining the lowest possible temps in a hot Texas environment. I also know it seems redundant to talk about restriction when I'm considering running 3-4 480 rads and 4 GPU blocks. I'm just trying to decide on the BEST possible performance, whether it's a balance of flow and cooling, or just a hardcore, low deltas and high flow system. I'm looking to OC a 3960X to around 5.4ghz and at least 3/4 of the gpus to around 1500-1800 mhz and want the cooling to be able to handle the temps that are going to be pushed.
I also do see your point on the Black ICE GTX 480's (based on the 360 tests that Martin did) Since the 360 is slowly scaling up with each 400rpm increase going from 1000-1400 by 58, then another 61 with the next increase to 1800, then 63 with the next increase to 2200 rpms. However the Alphacool Nexxxos UT60 (while each increase in RPM results in a slightly lower increase in heat dissipated) is a very close competitor. 1000-1400 increase of 59, 1400-1800 increase of 58, 1800-2200 increase of 57. While the heat dissipation numbers are going up by less and less, the prior numbers (600 and 1000RPM) were MUCH higher than the GTX 360. Thus, the Alphacool top end MIGHT not be as high as the GTX 360, however it has a much better range from low to high rpms in my opinion. Coincidently, It's also roughly $30 (USD) cheaper than the Black ICE when it comes to the 480's. I'm not so much concerned with that last part, just worth mentioning.
I also did notice that with a 325 W of dissipation at 2200 rpms w/ the 360, it would be at about 435w @ 10c Delta w/ 2200 rpm fans. So lets assume at it's current rate, that if i went full force with just the 4200rpm Gentle Typhoons that it would only go up to about....lets say an even 550w of dissipation (lower than the math actually suggests of course). Based on that 2x 480's would be at about 1100w of dissipation @ a 10c delta. Now...I know that adding a 3rd Rad will actually only give it about a 300w of dissipation increase since I have discovered that at 3 radiators diminishing returns starts to actually affect the loop by a considerate amount. So we'll say 1400w of of dissipation @ 10c @ 4200RPM. Here's where I run into trouble. How do I figure out the ~5c delta
Also...by learning more about this stuff, I've managed to make this more confusing on myself.
Side note: It seems that using compression fittings lowers my delta even further? I'm so confused haha.Edited by Flex85 - 4/16/12 at 3:27pm