Parallel Loop Theory

[Napalm360]

So Ivebeen messing around with paint looking at potential tubing routes

and It crossed my mind if you can have 3 gpus under a parallel block:

Spoiler

Which Basically has an inlet and an outlet like a cpu block but the fluid runs through multiple components that are connected together and this works fine.

 

BUT

 

 My Theory Loop:

Spoiler

Given that it works on the same principle as the 3x gpu block above when you have an inlet from a radiator going through components then out to another radiator, Will it work? Ty for your thoughts and looking forward hearing them

[SquintyG33Rs]

so short answer.... NO

 

it works with the GPU blocks because they are identical therefore they oppose the same resistance to waterflow. voltage and water are great analogues to eachother, voltage always takes the path of least resistance and if all paths have resistance it splits inverse proportionally to that resistance. aka it puts more V where there is lower R but a little still goes where resistance is high.

 

so in this case the flow will not be equal in GPU and CPU and GPU block has more stuff and is therefore more restrictive and will get less flow. it will be less cooled.

 

The GPU temp also matters more than the CPU because CPU clock won't change for a few degrees but GPU's will, therefore affecting performance.

[emosun]

i dont get how that 3 gpu setup would work without gpu 2 and 3 getting less water than 1.

[airdeano]

assumption is that the restriction level in the GPUs are equal. the restriction level of the CPU block vs the GPU block is not equal. like air, the coolant will travel a path of least resistance.if the CPU block is more restrictive than the GPU, the coolant will not flow to the CPU black and travel from the GPU (less restrictive) and return to the res instead of the CPU.

your 2nd pic will not cool and will bypass the GPU/CPU. plus the CPU blocks are usually designated with an inlet/outlet.

[SquintyG33Rs]

2 minutes ago, emosun said:

i dont get how that 3 gpu setup would work without gpu 2 and 3 getting less water than 1.

because physics

[emosun]

1 minute ago, SquintyG33Rs said:

because physics

maybe if the outlet tubes weren't all connected to each other and the inlet was on the center gpu.

otherwise the outlet of gpu 1 would push backward into 2 and 3 and the water flow in 2 and 3 would be less.

[SquintyG33Rs]

Just now, emosun said:

maybe if the outlet tubes weren't all connected to each other and the inlet was on the center gpu.

otherwise the outlet of gpu 1 would push backward into 2 and 3 and the water flow in 2 and 3 would be less.

no that's just not how it works. you can read the post right after you or my first answer for the explanation

[GirlFromYonder]

13 minutes ago, Napalm360 said:

So Ivebeen messing around with paint looking at potential tubing routes

and It crossed my mind if you can have 3 gpus under a parallel block:

  Hide contents

Which Basically has an inlet and an outlet like a cpu block but the fluid runs through multiple components that are connected together and this works fine.

 

BUT

 

 My Theory Loop:

  Hide contents

Given that it works on the same principle as the 3x gpu block above when you have an inlet from a radiator going through components then out to another radiator, Will it work? Ty for your thoughts and looking forward hearing them

the resistance would be a bitch to overcome.

[emosun]

1 minute ago, SquintyG33Rs said:

no that's just not how it works. you can read the post right after you or my first answer for the explanation

yeah and the path of least resistance is gpu 1.

gpu 2 and 3 are farther away and have gpu 1's outlet pushing against their outlets. maybe if they all had their own outlets it could work , or if the outlet was in gpu 3 near the bottom it might. but yeah no what he drew would work kinda crappy.

[SquintyG33Rs]

Just now, emosun said:

yeah and the path of least resistance is gpu 1.

gpu 2 and 3 are farther away and have gpu 1's outlet pushing against their outlets. maybe if they all had their own outlets it could work , or if the outlet was in gpu 3 near the bottom it might. but yeah no what he drew would work kinda crappy.

no because the blocks are identical (they all have the exact same resistance).... and continuous tube is 0 resistance (no funnel). so the flow splits evenly in all 3. they each get 1/3 of the flow

[W-L]

22 minutes ago, Napalm360 said:

-SNIP-

As said the first one works because they have equal resistance and flow rate through the blocks, the second setup with different lengths of tubing and resistance of different blocks and rads in this case will not work, the fluid will bypass the whole portion and just go through the CPU mainly. 

[emosun]

1 minute ago, SquintyG33Rs said:

no because the blocks are identical (they all have the exact same resistance).... and continuous tube is 0 resistance (no funnel). so the flow splits evenly in all 3. they each get 1/3 of the flow

which would be true if they didn't all share the same outlet. that's why rads inlet and outlet are kiddy corner to each other or end to end and not side by side , otherwise it wouldn't use the radiator  and simply go in and out

[emosun]

3 minutes ago, W-L said:

As said the first one works because they have equal resistance and flow rate through the blocks

card 1 has less resistance than 2 and 3 because card 1 outlet is pushing back on 2 and 3

if it exited out of card 3 it might work better.

[SquintyG33Rs]

Just now, emosun said:

which would be true if they didn't all share the same outlet. that's why rads inlet and outlet are kiddy corner to each other or end to end and not side by side , otherwise it wouldn't use the radiator  and simply go in and out

no there ... there is no path.... all the tubes on one side on inlets and all the ones on the other are outlets... and the only way between the two is going through the whole block of at least 1 GPU. and since they all have the same block and same resistance the flow splits equally in all of them. your intuition is not physics, sometimes physics are intuitive, this is not one of those cases for you.

[SquintyG33Rs]

Just now, emosun said:

card 1 has less resistance than 2 and 3 because card 1 outlet is pushing back on 2 and 3

if it exited out of card 3 it might work better.

if it worked the way you think it works it wouldn't even work with a single card on it's own or with cards in series

[W-L]

Just now, emosun said:

card 1 has less resistance than 2 and 3 because card 1 outlet is pushing back on 2 and 3

if it exited out of card 3 it might work better.

Semantically yes since that is the shortest path with the least frictional loss through flow but in reality the difference is so small it would be within margin of error, if the flow path had both inlets at top and bottom of the block stack with exactly the same length of tubing splitting into the top and bnottom then the flow would be perfectly balanced. 

[Napalm360]

Oh Lord what have I started, makes sense I think I need sleep also completely forgot about parallel circuits when thinking of this

 

Thanks for the activity and think of parallel GPUs block to be like a funnel with 3 heads, it evenly splits liquid into 3 containers

 

 

[emosun]

Just now, W-L said:

Semantically yes since that is the shortest path with the least frictional loss through flow but in reality the difference is so small it would be within margin of error

idk i think through sheer luck of "water moving around" would be the only thing keeping gpu 2 and 3 cooled.

 

2 minutes ago, W-L said:

if the flow path had both inlets at top and bottom of the block stack with exactly the same length of tubing splitting into the top and bnottom then the flow would be perfectly balanced

I agree with that.

it's the fact that the inlet and outlet are both card number 1 that makes 2 and 3 a dead zone sort of speaking.

[SquintyG33Rs]

i give up....

[Napalm360]

1 minute ago, emosun said:

idk i think through sheer luck of "water moving around" would be the only thing keeping gpu 2 and 3 cooled.

 

I agree with that.

it's the fact that the inlet and outlet are both card number 1 that makes 2 and 3 a dead zone sort of speaking.

Lol Its one of those things you take our word on it, As squinty said and I agree, fluid dynamics and electricity are extremely similar, one can describe the other

[Napalm360]

1 minute ago, SquintyG33Rs said:

i give up....

To quote Barnacules Nedrgasm, If you give someone Factual Points and even give a way to personally verify said point then education has has failed you

[W-L]

7 minutes ago, emosun said:

idk i think through sheer luck of "water moving around" would be the only thing keeping gpu 2 and 3 cooled.

I agree with that.
it's the fact that the inlet and outlet are both card number 1 that makes 2 and 3 a dead zone sort of speaking.

5 minutes ago, SquintyG33Rs said:

i give up....

The factor that would make it see a difference in flow is the extra distance away from the inlet and outlet of the lowest block but the losses with friction is very limited. The pressure applied by the pump will easily force fluid through all three blocks, not 100% evenly but very close. You will not have any kind of dead zone. 

[Napalm360]

@SquintyG33Rs so apply the same principle to a CPU and ram block, where there's a t fitting on each port of the ram block to allow the two tubes going to and from the CPU, would that work?

[emosun]

6 minutes ago, Napalm360 said:

Lol Its one of those things you take our word on it, As squinty said and I agree, fluid dynamics and electricity are extremely similar, one can describe the other

hmm ok

 

what happens when you split one line into 3 lines and then combine them again? where does the water or electricity go? the path of least resistance right?

if all line are the same length , then all get equal flow of electricity or water.

if they are different lengths like card 2 and card 3 being FARTHER away from the inlet and outlet , then they will receive less current.

[SquintyG33Rs]

2 minutes ago, Napalm360 said:

To quote Barnacules Nedrgasm, If you give someone Factual Points and even give a way to personally verify said point then education has has failed you

yeah. but psychology also taught us that in a position of equals it is near impossible to convince somebody of their mistake because of confirmation bias and every attempt to prove them wrong will only reinforce their convictions because we seem desperate to bring them to our side instead of just agreeing with them... humans are so dumb sometimes i swear.