Hey guys . I have two rtx 3090 and I was wondering if its possible to make games like Control / Quantum break use sli/nvlink using nvidia inspector.
I tried last night for a few hours but failed , the profiles i found online for quantum break didnt work .
did anyone here try ?

Thank you for your answer but thats not what i mean . I mean activating sli via nvidia inspector . I posted in this topic because I found an older topic about Red Dead Redemption 2 , and ppl did manage to get sli with 3090 .
 

 
 
 

 

So what you are saying is I can enable 4k sli on 2 x 2070 , but not on 2 x 3090 ? for quantum break / dx11 ?

wharrus5
So what you are saying is I can enable 4k sli on 2 x 2070 , but not on 2 x 3090 ? for quantum break / dx11 ?

Correct, but you’ll need to use drivers older than 2021. Even then it might not work though depending on game.

Nope, because dx11 profiles are now being ignored on 30 series cards.

wharrus5
guys i kept trying without luck . is there anyone who was able to use a sli profile dx11 in nvidia inspector with sli 3090 ? 

No, did you think we were just making it up? I posted a link directly from Nvidia themselves stating they have dropped SLI support for any and all DX11 titles on all 30 series cards going forward. 

wharrus5
nvidia also stated they limited the mining hashrate of 3060 and look how that turned out.
maybe something similar can be done here

The hashrate limiter on the 3060 wasn't at all "hacked". Nvidia mistakenly released a beta driver that didn't have the limiter in place. That is not even remotely the same situation. Nvidia themselves are NO LONGER actively developing SLI profiles, period. So there is nothing for them to "leak" with support for SLI on older games. They actually don't exist to leak for the new cards. 

Brad_Hawthorne
Nvidia made SLI a dumpster fire for 3000-series. Not sure what they were thinking.

Same thing game developers have been thinking for years.........."why bother wasting dev time and money on something almost no one uses?". 
 
Don't get me wrong I really like the idea of SLI and it's potential. But also as an avid user of it for literally every generation since it was possible, the pros just don't out weigh the cons. There are too many issues with frame pacing, negative scaling in some games, it just flat not working sometimes and all manner of other "quirks", it was only a matter of time before Nvidia themselves pulled the plug. 

Brad_Hawthorne

CraptacularOne

Brad_Hawthorne
Nvidia made SLI a dumpster fire for 3000-series. Not sure what they were thinking.

Same thing game developers have been thinking for years.........."why bother wasting dev time and money on something almost no one uses?". 
 
Don't get me wrong I really like the idea of SLI and it's potential. But also as an avid user of it for literally every generation since it was possible, the pros just don't out weigh the cons. There are too many issues with frame pacing, negative scaling in some games, it just flat not working sometimes and all manner of other "quirks", it was only a matter of time before Nvidia themselves pulled the plug. 

That's a chicken vs egg argument. Nvidia also couldn't be bothered to make it work right because devs wouldn't properly implement it. In an ideal world, SLI would have been a high speed bus interface that allowed GPU chiplets to talk to each other as though they were all a single GPU over several cards. Well, we all know that never materialized. I don't think Nvidia would want that anyways, because then people would be using 4-6 mid range cards to obliterate the premium cards. SLI never scaled well to begin with. That's not the fault of SLI as a concept, but SLI as a reality. I'd take a bet SLI would be designed a lot differently in 2022, than in 2004.

Well of course it would more than likely be designed and developed differently in the future. That's just a hypothetical argument however with no more or less merit than the chicken or the egg. Neither of us can change the way it happened or was developed. We can only hope that should Nvidia or AMD want to develop a better method that they learn from what didn't work. But I suspect this is the last we've seen of consumer grade SLI for quite a while at least with multiple PCB cards. 
 
As I said I was a big fan of and supporter of SLI and Crossfire for that matter and have used them both in almost every iteration they have been possible. However in the era of $1500 flagship GPUs it's really just not feasible to think the majority of users would be keen to or even be able to afford $3k investment for games. The real area where SLI and Crossfire should have shined was in the lower end segment of the market to bolster GPU performance when needed or more accurately, could be afforded. Most people only spend $200-$300 on a gaming GPU (as evidenced by the steam hardware surveys) and historically the "60" series class cards have dominated the market volume. However SLI and Crossfire scaled very poorly in the lower to midrange GPUs and what people learned was that buying a second $200+ card and then adding it to their system often resulted in at best a 50% performance gain. Hence was born the sage advice of "don't bother with SLI on lower end cards just buy a single faster GPU". That's what killed SLI and Crossfire. 
 
Sure AMD and Nvidia could have done more to directly support developers (and even did on both sides) to add or help implement multiGPU support for their games. But the damage was done at that point, from very poor frame pacing (the infamous "microstutter") to poor performance scaling, to poor temperatures and added cost. I'm quite frankly surprised it lasted this long and isn't completely officially dead. 

Would be interesting to see if this works