If I understand correctly, in a series GPU setup you get the same flow rate in both gpu blocks, but the temp is usually a little higher in the second(EX- GPU1 *47c, GPU2 *49c). In parallel the flow rate is the same in both blocks, although reduced by half compared to series, and both cards meet in the middle of the temp (EX- GPU1 *48c, GPU2 *48c). Does that sound correct?
In the loop I'm planning I am going to have a high restriction since it's 2gpu's, a cpu, and 2 rads. The biggest choke point is currently the CPU block, but I already have that and it matches color scheme so well, so I'm looking to reduce restrictiveness where I can.
My loop will be Res>MCP35x>Swiftech Apogee HD>EK XTX240>EK FC580GTX>EK FC580GTX>EK XTX360>res.
From what I've read the MCP35x is made for this kind of loop with high restriction (please correct me if I'm wrong).
The main question I have is parallel or series for the GPU's?
ok i have already done this and i have the test results to show the flow rates on page five of the build i ended up going with a total parallel loop for the cpu, NB, 2nd gpu and mosfet 1st gpu on the other line of the parallel loop but have a look at my build for your self
Sweet, thanks Damn that is a sexy rad setup on the second loop. So you just put a Y to split the line, and have both lines going into GPU1 and only one line going to GPU2 from GPU1? I can't think of a way that would work in my setup without using way too many 90's. It definitely got me thinking of different ways to set up my loop tho. It's got me really thinking about how to implement a rad box I've got this micro case that I got for free from newegg that I could mount to the bottom of my desk. But i'm thinking that I'd have to go dual loop for that or a very powerful pump. Or would I be able to put a second pump in the rad box without adding a second loop? Now I've gotten myself thinking too much again lol, time for many more hours of research!
what i found was this, for every part you add to the loop on a series loop the flow rate would drop for each part. so i paralleled the entire loop so that i could achive the highest flow rate through the pump the best results would be if i 3 wayed the loop then i would have gotton the pump to full speed but would have slowed each lane down to 2000 or less, and i would like to keep it above that. right now it is at 5000 so each lane is at 2500 and the temps are cool. one 655 pump can handle 4 rads and 15 blocks in this type of configuration, i do not know what the limit is i have not hit it yet. i went with two loops because of the restrictive nature the blocks and rads attched to that loop.