• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Fidelity Super Resolution in 2021

Associate
Joined
20 Nov 2020
Posts
1,120
Quite a bit of stuff in the DLSS and RT pipelines don't use Tensor cores (even though originally claimed was the case) even though in a best case scenario they can provide a significant performance boost over shaders due to the lack of flexibility and having to hand tune to avoid penalities i.e. Tensor processing is often last in the chain and can't be run simultaneous with other work.

When and where you can use Tensor cores though the speed up can be significant.

At lower resolutions you could probably get away with using Shaders but towards 4+K resolutions you would have to rely on Tensor cores to get decent performance gains.

Off the top of my head IIRC doing it on the shaders on Ampere takes about 4ms at 1080p compared to ~1ms on the Tensor cores at 1440p (output resolution) - I don't know the numbers for 1080p on Tensor but I'd assume sub 1ms.
I am assuming this is still the case (that they are not doing too much work in the gaming cards ) but i was talking about this particular case where it is known the tensor cores were not used. It was still good enough performance. And yet some people are convinced it can't be done without the tensor cores, even if most of them have used DLSS done without tensor cores already and they were impressed at that time. :)
 
Permabanned
Joined
21 Feb 2021
Posts
474
tensor cores are probably a huge hoax

its probably a clever way to segment their "brand new" gpus from older ones

i clearly lost faith in dlss, after seeing how FSR does a great job at 4k with quality and ultra quality modes, just like DLSS...
 
Associate
Joined
20 Nov 2020
Posts
1,120
tensor cores are probably a huge hoax

its probably a clever way to segment their "brand new" gpus from older ones

i clearly lost faith in dlss, after seeing how FSR does a great job at 4k with quality and ultra quality modes, just like DLSS...
It is easier to fool people at 4k, not too many of us can see big differences between 1440p and 4k, that's why both features have the best results when the upscaled res is 4k and they lose quality as the resolution gets lower.

I see Dota came right away with FSR support this is huge for AMD, too bad they can't get into UE that easy. God knows what contracts Epic has, they are Sony's slaves on consoles and Nvidia's slaves on PC's. If FSR gets into Fortnite too, then no one will use DLSS in that game. :)
 
Man of Honour
Joined
13 Oct 2006
Posts
91,687
tensor cores are probably a huge hoax

its probably a clever way to segment their "brand new" gpus from older ones

i clearly lost faith in dlss, after seeing how FSR does a great job at 4k with quality and ultra quality modes, just like DLSS...

Well aside from the fact DLSS or a good TAAU implementation gives a far higher FPS up lift at the same quality level as FSR...

As for Tensor cores being a hoax #facepalm#
 
Soldato
Joined
18 Feb 2015
Posts
6,489
0bpkjcsmh4771.jpg

https://www.dota2.com/newsentry/2992060508110412253
 
Caporegime
Joined
12 Jul 2007
Posts
40,878
Location
United Kingdom
And zen you see this:
https://www.reddit.com/r/Amd/comments/o6skjq/digital_foundry_made_a_critical_mistake_with/
How was Alex shilling? He is the best upscaling expert you can find on internet.
Lol another example to go with the ones already posted.

They messed up Godfall also by enabling sharpening, then compared to native and said it looks worse at Ultra Quality.

That’s why HUB looks so different to their video.

DF just don’t have any credibility anymore since they took the ££ and posted that sponsored content days before official reviews could be posted.

Dota? Next it will be cs:source. It's 2007 all over again
Made me chuckle...
Quake II (RTX) is a first-person shooter video game first released in December 1997.
 
Last edited:
Soldato
OP
Joined
6 Feb 2019
Posts
17,857
It maybe fake ... https://www.reddit.com/r/Amd/comments/o6rfqi/fsr_on_dota_2_is_amazing_lowest_render_resolution/

I suspected as the trees in the background appear to have more detail than source. That's not what FSR does.

That particular comparison seems fake or otherwise misleading intentionally - as you mention the FSR on image contains additional tree branches and detail that is gone with FSR off and that should be impossible all else being equal. All the Reddit comments are calling him out on it too, there is no way those two images are using the same internal resolution - it's possible he made a mistake or that FSR in that game is a bit wonky and actually runs at a higher resolution than you'd think
 
Last edited:
Man of Honour
Joined
13 Oct 2006
Posts
91,687
That particular comparison seems fake or otherwise misleading intentionally - as you mention the FSR on image contains additional tree branches and detail that is gone with FSR off and that should be impossible all else being equal. All the Reddit comments are calling him out on it too, there is no way those two images are using the same internal resolution - it's possible he made a mistake or that FSR in that game is a bit wonky and actually runs at a higher resolution than you'd think

Looks to me someone is comparing the lower resolution through normal in-game resolution scaling to native res (top image) and FSR output to native res (bottom image) either that or they are deliberately being mislead - some people seem to find that amusing to do.

The top image looks like what happens when you render at 60-65% resolution scale and upscale via a bilinear filter to native resolution, the bottom ones looks like around 80% resolution scale filtered through FSR.

EDIT: I see they've corrected themselves - the top image is in-game resolution scaling from 40% to native the bottom image is 50% resolution scaling via FSR - though even that looks a bit off to me but then they've cropped and scaled, etc. the image which somewhat obscures the original and native resolution. Using FSR on and off in that context is just plain misleading but I don't really expect any better.
 
Last edited:
Associate
Joined
20 Nov 2020
Posts
1,120
Why would I? TAAU in Quake 2 RTX provides upto 3x the performance uplift at the same quality level as FSR.
But how do you know how the FSR will look if you don't have it? Look at that reddit post where someone is challenging Alex, FSR looks arguably better than TAAU.
I was joking anyway but i said in the past FSR is like a cheap chinese tech, Maybe it is not as good as the western tech but it will conquer the world. :D
 
Man of Honour
Joined
13 Oct 2006
Posts
91,687
But how do you know how the FSR will look if you don't have it? Look at that reddit post where someone is challenging Alex, FSR looks arguably better than TAAU.
I was joking anyway but i said in the past FSR is like a cheap chinese tech, Maybe it is not as good as the western tech but it will conquer the world. :D

With enough experience you can compare the before and after images and have a pretty good idea of the quality comparison (plus stuff like a quantitative difference analysis on the screenshots).

Unreal TAAU IMO is not one of the better implementations - the latest experimental version has some big improvements but it still needs a lot of work IMO.
 
Last edited:
Associate
Joined
20 Nov 2020
Posts
1,120
We can say whatever we want but adding FSR to DOTA2 was a great idea from AMD, it is like they hired Jensen to make their strategy. That game has tens of times more users than any DLSS game at any time. And FSR will make a lot of them happy.
 
Soldato
Joined
18 Feb 2015
Posts
6,489
It maybe fake ... https://www.reddit.com/r/Amd/comments/o6rfqi/fsr_on_dota_2_is_amazing_lowest_render_resolution/

I suspected as the trees in the background appear to have more detail than source. That's not what FSR does.

That particular comparison seems fake or otherwise misleading intentionally - as you mention the FSR on image contains additional tree branches and detail that is gone with FSR off and that should be impossible all else being equal. All the Reddit comments are calling him out on it too, there is no way those two images are using the same internal resolution - it's possible he made a mistake or that FSR in that game is a bit wonky and actually runs at a higher resolution than you'd think

Looks to me someone is comparing the lower resolution through normal in-game resolution scaling to native res (top image) and FSR output to native res (bottom image) either that or they are deliberately being mislead - some people seem to find that amusing to do.

The top image looks like what happens when you render at 60-65% resolution scale and upscale via a bilinear filter to native resolution, the bottom ones looks like around 80% resolution scale filtered through FSR.

EDIT: I see they've corrected themselves - the top image is in-game resolution scaling from 40% to native the bottom image is 50% resolution scaling via FSR - though even that looks a bit off to me but then they've cropped and scaled, etc. the image which somewhat obscures the original and native resolution. Using FSR on and off in that context is just plain misleading but I don't really expect any better.

It's really simple, it's res scale vs FSR, both 50%.

https://i.imgur.com/ugCmdqv.jpg
https://i.imgur.com/Ni7gcjz.jpg
 
Back
Top Bottom