30
AMD plans for FSR4 to be fully AI-based — designed to improve quality and maximize power efficiency
(www.tomshardware.com)
this post was submitted on 14 Sep 2024
30 points (85.7% liked)
Games
16785 readers
811 users here now
Video game news oriented community. No NanoUFO is not a bot :)
Posts.
- News oriented content (general reviews, previews or retrospectives allowed).
- Broad discussion posts (preferably not only about a specific game).
- No humor/memes etc..
- No affiliate links
- No advertising.
- No clickbait, editorialized, sensational titles. State the game in question in the title. No all caps.
- No self promotion.
- No duplicate posts, newer post will be deleted unless there is more discussion in one of the posts.
- No politics.
Comments.
- No personal attacks.
- Obey instance rules.
- No low effort comments(one or two words, emoji etc..)
- Please use spoiler tags for spoilers.
My goal is just to have a community where people can go and see what new game news is out for the day and comment on it.
Other communities:
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
AI image upscaleing isn't something I would associate with being energy efficient or fast. I wonder how that's supposed to work?
It seems like it would be extremely fast to me. Take a 50x50 block of pixels and expand those across a 100x100 pixel grid leaving blank pixels were you have missing data. If a blank pixel is surrounded by blue pixels, the probability of the missing pixel being blue is fairly high, I would assume.
That is a problem that is perfect for AI, actually. There is an actual algorithm that can be used for upscaling, but at its core, its likely boiled down to a single function and AI's are excellent for replicating the output of basic functions. It's not a perfect result, but it's tolerable.
If this example is correct or not for FSR, I have no clue. However, having AI shit out data based on a probability is mostly what they do.
I'm very much not an expert, but I'd imagine it's similar to how AES-NI works: the task is CPU/GPU-intensive until specific instructions are designed to do whatever blackmagicfuckery level math is required, and once it's in hardware it's more both power efficient and faster.
Without more detail we can only assume, but I would imagine it working the same way that DLSS is (presumed?) to work.
Most of the upscaling is done by their TAA algorithm that's a part of FSR3.1, then the image will be cleaned up with their "AI" component for more image stability.