cancel
Showing results for 
Search instead for 
Did you mean: 

Question’ two gpu’s in same rig.

Rob_W_
Level 12
Is there a problem in having a Titan v and a 2080ti in the rig at the same time, ?
*reason I ask is I would like to bench both cards individually on the same platform but without having to swap the loop to do it.
I think pcie slots 1 and 2 both work at 16x? So I could just use switches on mobo to disable the one not being tested, to my knowledge I think both slots should give same performance but please correct me if I am wrong!
It looks also that both cards would use the same driver version.
Any thoughts on this?
2,178 Views
18 REPLIES 18

F0x135
Level 7
Got to make sure the mobo you're going with is going to have enough space. I'm unsure of the clearance with TitanV, but the 2080Ti is 2.75 slot. I wanted to replace one of my 1080Ti's with a 2080Ti, but it won't fit in this z370-F setup. The x2 1080Ti's are already kissing with a 2.5 slot clearance.

Korth
Level 14
I'm not as certain that both x16 slots would give identical performance.

One of them is simply closer to the CPU than the other. It has shorter physical and electrical lengths. Which might translate into lower signal losses or latencies. So I expect that ASUS may have hardwired or firmcoded this slot in a more preferential way ... after all, multi-GPU setups usually put most of the load on the primary GPU anyhow.

But then again ... you can use PCIe extenders and risers, apparently with surprising lengths, and still have little measurable loss in actual performance.

The most comprehensive way to answer this question would be to test both cards, then swap and test again ... lol, the answer would be interesting but this approach sort of defeats the reason you even ask the question.
"All opinions are not equal. Some are a very great deal more robust, sophisticated and well supported in logic and argument than others." - Douglas Adams

[/Korth]

Korth wrote:
I'm not as certain that both x16 slots would give identical performance.

One of them is simply closer to the CPU than the other. It has shorter physical and electrical lengths. Which might translate into lower signal losses or latencies. So I expect that ASUS may have hardwired or firmcoded this slot in a more preferential way ... after all, multi-GPU setups usually put most of the load on the primary GPU anyhow.

But then again ... you can use PCIe extenders and risers, apparently with surprising lengths, and still have little measurable loss in actual performance.

The most comprehensive way to answer this question would be to test both cards, then swap and test again ... lol, the answer would be interesting but this approach sort of defeats the reason you even ask the question.


that’s what my concern was Korth, wether there would be degradation between slots.
Still as I’m waiting for the Titan to be repaired and I haven’t got the 2080ti yet so it’s all a bit hypothetical.
I agree it would be interesting to swap them between slots ( a lot of work ) to compare but was hoping to avoide that scenario,
Still you have aroused my interest now so I will probably try it just to see!*

Menthol
Level 14
Yes, I have done the same thing with different model cards

Rob_W_
Level 12
Both will have EK water blocks , Titan on water is single slot, 2080ti should be reduced to that as well, it’s the stock coolers that boost there size, as to the case**airflow and space not a problem ( obsidian 900d case) .*

Korth
Level 14
For reference, I submit Linus's experiment ... the TL;DR summary is that he plugged extenders into extenders (including some "janky old" extenders of dubious reliability) until he hit stability limits and crippling fps losses somewhere around 3.0 metres ... so apparently a few centimeters makes little difference ... but that still doesn't rule out motherboard/firmware optimization choices, I don't think ASUS implements a "T-Topology" counterpart into their PCIe slots.
"All opinions are not equal. Some are a very great deal more robust, sophisticated and well supported in logic and argument than others." - Douglas Adams

[/Korth]

Silent_Scone
Super Moderator
For sake of consistency, I would benchmark both of them independently.
13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

Silent Scone wrote:
For sake of consistency, I would benchmark both of them independently.


I assumed that was implied

Menthol wrote:
I assumed that was implied


But in the same slot 😉
13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090