ramthor's monster build

Yeah what i meant is you could manage that with a couple of 980tis, so 2 titans aint gonna be stressing out much, i guess it means you can run them without an over clock for a while until games get a little more demanding in a few years then.
 
BluD said:
Yeah what i meant is you could manage that with a couple of 980tis, so 2 titans aint gonna be stressing out much, i guess it means you can run them without an over clock for a while until games get a little more demanding in a few years then.
Click to expand...

That is not true. Look at this video. The 980Ti's struggle to stay around 60 average with minimums as low as 40FPS noticeable especially when he goes through the grass.

https://www.youtube.com/watch?v=j9komRz3vuA
 
A pair of overclocked GTX980Ti's are good enough for 4K.

That said they're probably on the obsolete driver list now, so maybe they are starting to struggle with the latest games. I know my 780ti isn't doing great nowadays.
 
EVH said:
11 posts above too :o. If you want to change the title let me know.
Click to expand...

Yeah, the title no longer reflects the price or specs. I will probably start a new thread with the updated specs and final cost when I'll bloody make up my mind and start ordering things. Just change it to Monster build for now. Thanks!
 
ramthor said:
That is not true. Look at this video. The 980Ti's struggle to stay around 60 average with minimums as low as 40FPS noticeable especially when he goes through the grass.

https://www.youtube.com/watch?v=j9komRz3vuA
Click to expand...

Yeah thats true, ultra grass really kills the fps, if i switch from vhigh to ultra i lose 20-30 fps depending on the area. I guess if ya spending £8k ya would wanna use ultra grass :)
 
ramthor said:
I looked at this diagram when I thought I need to place the 2nd card in the third slot to get x16 x16

2ca5712050.jpg


Is the distance between those slots 80mm? I can't see any EK-FC Terminal with that size.

I Also plan to use flexible tube in the lower chamber as it's out of sight and doesn't matter.

Thanks for the tip that I can't use those multiports for what I planned. That was the only reason I went for that rad.

EK is in the process of coming with their own SLI HD bridge. I think I can use 2 flexible regular bridges while waiting.
Click to expand...

It is the 3rd x16 slot, yes... but the distance between the 1st slot and second slot is double-width... so you have:

1 slot = pci-e x16
2 slot = empty
3 slot = pci-e x16 (size slot, but only works at x8 speed)
4 slot = pci-e x8
5 slot = pci-e x16 (the one you want to put second card in as per diagram to get full x16 speed)

So that's 4 slot spacing which is 80mm.

I double checked the size of my EVGA HB SLI bridge before mentioning that to make sure I wasn't giving you duff information - definitely 80mm.


And yes, you can use 2x flex bridges while you're waiting - that's what I've been doing.
 
crinkleshoes said:
It is the 3rd x16 slot, yes... but the distance between the 1st slot and second slot is double-width... so you have:

1 slot = pci-e x16
2 slot = empty
3 slot = pci-e x16 (size slot, but only works at x8 speed)
4 slot = pci-e x8
5 slot = pci-e x16 (the one you want to put second card in as per diagram to get full x16 speed)

So that's 4 slot spacing which is 80mm.

I double checked the size of my EVGA HB SLI bridge before mentioning that to make sure I wasn't giving you duff information - definitely 80mm.


And yes, you can use 2x flex bridges while you're waiting - that's what I've been doing.
Click to expand...

Thank you for confirmation. Now I know I can't use a FC Terminal and I have to build the connection between GPU blocks myself.

This arises new questions. :)

I'm sticking with EK for all water cooling connections for increased compatibility. So for GPU block connectivity their website lists HD connectors and pre-cut hard tubing. My questions are:

1. Why HD fittings and not HDC as the rest of the loop?

2. Is it worth buying the pre-cut tubes? Can't I cut them myself? Why would they sell the pre-cut tubes? Is it difficult to get them cut in the same size or what?
 
The HD fittings are designed for this kind of connection where the pressure between the GPUs helps keep the connections together. You can use either, though.

I bought pre-cut tubes - but only because they were so cheap and I knew that they would fit perfectly rather than taking extra time to get it just right. It would be easy enough to cut them yourself.

I wanted an FC terminal too... but like you, couldn't find one the right size.
 
I read that parallel is superior to serial as it is less restrictive on the loop flow.

Am I right in assuming if I want to go for a parallel configuration I need a total of 4 HD fittings (between GPU blocks) and 2 HDC fittings for in/out flow?
 
That's correct yes.

I prefer serial, but that's a personal preference thing. You shouldn't have any trouble with flow with the pump you have - although if you are worried about restrictions, you could always get the EK Dual D5 pump like I have... it's two D5 pumps stuck together that work in serial.

Product name is "EK-XTOP Revo Dual D5 PWM Serial - (incl. 2x pump)"

I went for that pump due to having quite so many radiators.

How it works in my head is that in serial... they're both forced to full flow whereas in parallel you could get some lower flow / stagnant areas within the GPU blocks and I would rather have slightly lower total flow than potentially reduced cooling per gpu block. It might work better in real life than it does in my head... but that's my reasoning behind my preference.
 
Last edited:
crinkleshoes said:
That's correct yes.
Click to expand...

Thanks.

I prefer serial, but that's a personal preference thing. You shouldn't have any trouble with flow with the pump you have - although if you are worried about restrictions, you could always get the EK Dual D5 pump like I have... it's two D5 pumps stuck together that work in serial.

Product name is "EK-XTOP Revo Dual D5 PWM Serial - (incl. 2x pump)"

I went for that pump due to having quite so many radiators.
Click to expand...

I've considered getting the dual pump but if I want to split the loop in the future I'll end up with having 3 or 4 pumps for 2 radiators which is really unnecessary.

How it works in my head is that in serial... they're both forced to full flow whereas in parallel you could get some lower flow / stagnant areas within the GPU blocks and I would rather have slightly lower total flow than potentially reduced cooling per gpu block. It might work better in real life than it does in my head... but that's my reasoning behind my preference.
Click to expand...

That is how I imagined it too but in reality there is not a significant difference between serial and parallel just as the fact that the loop components order is not significant either.

Counter intuitive, I know, but that's what I read it was observed through experimentation.

So I think it comes down to aesthetics when choosing one or the other.
 
I've switched between serial and parallel loads of times, if there is a difference it must be fairly negligible as i dont recall any changes in temps, dont worry about dual pumps either, i have 4 rads 2 gpu and one CPU block and it runs fine on speed 2 or above using a single d5.
 
Fair play... I thought that might be the case... just my mind likes to work the way it does hehe.

I thought a single pump should be more than enough... I just over-specced for the sake of feeling better :)
 
original.wood said:
Are you forgetting altogether about the HTC Vive?
Click to expand...

:)
Thank you for noticing!

I have omitted it on purpose from the build because it has gotten ridiculously expensive already and I needed to protect what's left of my sanity. :D

I will still buy it. Just not right now. I have plenty of things that can go wrong without adding even more complexity to an already complex project.
 
You must log in or register to reply here.
Back
Top Bottom