CPU/GPU block flow direction. Does it really matter?

Colonel_Klinck

Colonel_Klinck

Colonel_Klinck

Soldato
Joined
3 Oct 2007
Posts
12,179
Location
London, UK
I know CPU and GPU blocks have in/out ports but does it really matter which way the coolant flows? I ask because if I could reverse the flow duration on my GPU block it would really help my tube layout.
 
For my EK blocks, with the GPU it doesn't make much difference, maybe 1 or 2 degrees max so go with whatever is easiest layout wise.

With the CPU block the jetplate design means you're adding turbulence and restricting flow if you swap the in and out. I imagine this is more of an issue if you have a low flow rate, high water flow can likely brute force this. Temperature difference will be higher than with the GPU, I can't give you a solid number though.

As String says, I think a good rule of thumb is that it matters for the CPU, and doesn't really matter for the GPU.
 
Both my blocks have in and out indicators, but i know that many do not.

My take is that if direction didnt matter then they wouldn't bother adding the expense of marking in and out flow on the product and in manuals. So if there are markings i would stick to them.

As for what will happen if you don't, it could simply be that the block is less efficient. I doubt it is any more than that. Unless a particular flow pattern on a gpu block stops hot spots accumulating in the block. Thats all i can think of for worst case but I am only guessing.

You could always try it and if everything works for you then stick with it.
 
Colonel_Klinck said:
I know CPU and GPU blocks have in/out ports but does it really matter which way the coolant flows? I ask because if I could reverse the flow duration on my GPU block it would really help my tube layout.
Click to expand...


Bykski for Zotac 1080ti Amp/ Amp Extreme. It was fine but now I've mounted the GC vertically to show off the GC block I'm going to have to change the routing. If I could reverse through GC it would but I don't want to lose any cooling efficiency.
 
The left channel is flow in.

IMG-5469.jpg
 
Similar to an EK GPU block, the difference will be 1 or 2 degrees at most on your core. It's only really down to when it hits the VRMs I think. If you use the inlet, it goes Core > VRM. If you use the outlet as the inlet it goes VRM > Core (I think... I maybe wrong, it's something like that anyway!). I wouldn't worry too much.
 
Parsley said:
Similar to an EK GPU block, the difference will be 1 or 2 degrees at most on your core. It's only really down to when it hits the VRMs I think. If you use the inlet, it goes Core > VRM. If you use the outlet as the inlet it goes VRM > Core (I think... I maybe wrong, it's something like that anyway!). I wouldn't worry too much.
Click to expand...

Cheers. I'll give it a go. I'm miles from temps being as issue. GPU under full load boost @2073ghz is 54c
 
Well I've switched to hard tubes, only cocked up a couple of bends which I'm guessing isn't too bad for a first attempt. Am pretty chuffed with result. As you all said hasn't made any real difference to temps, well maybe a couple of degrees higher, full load for 30 mins in Heaven Bench and GPU sat around 52c with a max of 56c. Had to tip the case around a bit to flush air from GC block but other than that all good. :D

IMG-5591.jpg
 
Temps update. After 90 mins of of playing BF5 the GPU hit a max of 65c and CPU 64c. Still ok considering its only a 240mm x 60mm rad for both. Thought I'd have to invest in a second rad. Maybe when summer comes around.
 
Distracted said:
As for what will happen if you don't, it could simply be that the block is less efficient.
Click to expand...
This is particularly true of the new EK blocks that use jet plates (like CPUs) to as a gimmick, when the block is plumbed in "backwards" they increase restriction while having no impact on performance (minus any that may/may not be lost due to added restriction or loss of flow).
 
You must log in or register to reply here.
Back
Top Bottom