How would this work in theory
>>108540713>How would this work in theorysteam ask AI and AI make it up
translation "i work for steam and added a gay function for normies, no i will not buy an ad corporate wont give me the cash amd i wont pay out of pocket"
Steam already tracks fps in your game. They can combine that with hardware survey and use it to estimate fps for given hardware configuration. It probably won't take into account graphics settings so it won't be as reliable I guess.
>>108540713>User has Processor X, GPU Y and # GBs of RAM>Steam checks the average FPS for that combination from the user data they have
I'm pretty sure steam's system of showing FPS in game takes note of your specs and your framerate.
>menu has 10000fps>game runs at 15>can't tell if dlss is on
>>108540713>How would this work in theoryI mean, it literally says right thereI assume they will be data mining FPS data from users, and then displaying the average from users with similar specs to you
>>108540713GOG chuddies keep winning desu
>>108540713Specs + existing stats.
>>108540752yeah but games have different graphics settings as well, don't they?
>>108540901It's just taking the FPS they have, then what specs that machine has, combine all FPS from that configuration or similar and give you an estimate. Settings don't matter at all if you're just making a combined estimate
>>108540901its basically the modern version of "can you run it" website. it doesn't care about what graphics settings you use. all that matters is your specs and framerate/frametime.
>>108540823>Will beSteam has a shit ton of telemetry already.
>>108540713They can record the fps when you play the game along with the other user statistics they record, like game time etc.
I wouldn't mind. I have a 1050ti and I bought a ton of games off of humble bundle over the years so I'm always wondering which game I can run at a cozy 30fps or which will just shit the bed on low graphics.
On the steam deck it asks if you want to send anonymous frame data I guess I finally know why they asked for that in the first place
>>108540713>How would this work in theorysteam steals your hardware and ingame data and then calculates and avg to the user. its not hard to imagine, it works super easy. they simply steal data. what you gonna do about it?
>>108540901it should give a rough estimate (eg: what resolution, was their frametime good, did they enable hdr, did they experience any crashes)
>>108541077>they simply steal dataHardware scanning is opt-in.
>>108540713aside from obvious spying>videya card benchmarking sites exist>ask (((corporate))) ai to play a new game on (((average))) hardware A>get fps X>retrieve fps of a popular game on (((average))) hardware A>get fps Y>retrieve fps of a popular game on local hardware B>get fps Z>Z/Y gives a ratio Q, which tells you exactly how better or worse B is than A>multiply X and Q, get required fpsgranted, A needs to be shit, or steam needs several gamer setups (poorfag, office, normie casual setup, rich, autism) as well as extensive coverage of videya cards, but me thinks it can be donetell me if my math is wrong, i've failed school
>>108541216>extensive coverage of videya cardsby benchmarking sites
>>108540713>WOAW HOW COULD THEY KNOW?????one fucking guy submits his benchmarks and you can do simple arithmetics to estimate everyone else's performance on his settings
>>108541216>multiply X and QX wouldn't make sense to use across different titles. You're retarded. They would just look for other players with similar hardware to yours running the game you want to buy
>>108541285wherever did you get an idea this one calculation is made for all games? it's for one game anon. this calculation needs to made anew for each title.now, one might think this is costly. however>(((corporate))) ai can be replaced by jeets>with a sufficient amount of testing gamer setups this way allows someone to request estimated fps before anyone plays a game>once the initial investment is made, there's only the cost of maintaining stuff (much cheaper)>if (((corporate))) ai is used, the fps bench marking allows many opportunities to teach and study neural networks, esp when in-house solutions are concerned>no risk of lawsuits and shit, pretty damn legaltho this is theoretical and the irl answer is spying. unfortunately.
>>108540713is this not literally what ai is good for?>good old multilayer perceptron>universal function approximator>dataset of specs matched with performance for some game>input specs>output predicted performance
>>108541354Why would you multiply X and Q then? You don't need to do that if you're already gathering fps data for hardware similar to B. Yes. You are retarded
>>108541354>spyingthere is tolerable telemetry and spying telemetryif valve logs your average framerate along with only e.g. your processor clock speed, then who carest. used to be schizo now only half schizo
>>108541215sure it is
>>108540713>Valve launches yet another invasive service that steals your dataEpic chads winning
>>108541395because Q is derived from Z, and Z is 1 local hardware, and people possess many different setups. so, for K*Z, where K is the amount of unique setups requesting fps data, you need K X*Q calculations. but there's 1 game, and it's tested once, after which all the Z's are scrapped (a cheap operation). you play the game once and get fps for all the hardware after a little bit of math>>108541424i do not trust any spying. first you get a little, then you get a lot.
>>108541471>first you get a little, then you get a lotthis is a function of how many straight white men are in control
i assume this is due to the amount of people who refund because games all run like shit now thanks to garbage engines like unreal flooding the store with low quality bullshit every single day and also thanks to imaginary, useless tech like dlss that doesn't work.it'd be the most useful if it was using opt-in data anonymously in a graph where you can browse & compare performance graphs from similar hardware to your own hardware, or this could just be implemented as an "avg FPS" field when you manually tick the field to include your PC specs when you leave a steam review, which would be next to useless as a feature, as i don't see any real value in now needing to comb through reviews looking for ones that only kinda/sorta match your own hardware.
>>108541513steam is trying to build a database of what runs good and what doesn't on their platform so they can eliminate the fake games and build a list of stuff that works on their deck/machine.
>>108540901yes, but it will even out when there's enough samples
>>108540713Doesn't seem particularly hard to do? it's just math
Still a DRM.
>>108540713It won't. Days when you could measure FPS and tell how the game performs are long gone. Now everybody runs games at 200 fps and random frame delay.
>How would this workframerate[cpu][gpu]maybe framerate[cpu][gpu][ram]
>>108540722They have telemetry from users. They don't need to guess.
>>108540822>I know more than Valve engineers
>>108540713Like this.You know, it'd be easier to go to the tweet to get the image full of tiny letters than to make a thread asking how it works.
how is this bad exactly?
>>108540713Probably something only for Steam Cube and Steam deck hardware.
Why do gamers care more about frames than polygons?
>>108542552It will hurt sales of corpo funded AAA slop that runs like shit on launch, please think of the billionaires!
>>108540713Oy vey, they better show the FPS at 240p upscaled to 4k with 6x framegen or an agent might have to teach Mr. Newell about noose tying.
>>108541279>gpu performance is the only metric that affects game fpsretard
>>108542598it's the metric that's going to be the determinant of it in any case where you're even asking about your performance
>>108540713I don't think it'll work very well, but we'll see. I consistently play games that /v/ claims are unplayable, and they work great on my machine. People will claim to have the same or better specs and that it barely runs, and almost without fail it's fine on my end. If these people are just lying for whatever reason, then maybe this will work. If these people have something else going on - they pirate everything so there are fifty coin miners running in the background, they 'built their own' without a clue and don't have XMP on, etc - then I would expect the same thing to happen here and skew the results.
>>108540901It's not meant to be precise. It'll give an approximate estimate that should be enough to give you an idea of whether your system can run it well enough to be worth playing.
>>108542569Because frame rate directly impacts controls and feedback. Even if you're just playing a single pixel video game where you just push a button when it changes from black to white, the FPS matters.
>>108541982he makes good points, they can detect dlss and generated frames but discarding frames from menus or any non gameplay will be a tough task
>>108541022What have you found that you can't play? I've upgraded from the 1050ti now because it was limiting me too much in Unreal Engine, but even that was only on big projects that I put together like a meathead idiot. For games I never found anything that was unplayable, only some things I had to turn down settings or maybe had some stutter or obvious dips at times. Overall it was really usable for even pretty new games, which is crazy when you think that it is basically considered ewaste now, and what you are supposed to replace it with costs a month's wages.
>>108540713Exactly as it's described, what the fuck are you even confused about?
>>108540901people will usually try to max the settings while keeping 60 fps or so (depends on monitor i guess but most people will use that, i understand some games like dota 2 or counter strike will target 100 or whatever but you get the idea). the outliers shouldn't matter when they have a big enough userbase
there are like 5 gamers with a blackwell 6000 tho?
>>108543858No it doesn't. You're not a cat with super reflexes. Humans can't react that fast.
>>108543903>I know more than Valve engineers and my babby's first superficial observations are sooooo smart I bet they didn't even think about it they should hire me
>>108544159framelet cope
>>108544216Ah another valve dick sucker. Never understood this.
>>108540730Most people never touch graphical settings any more.
>>108545343I always disable shadows.
>>108546283I hope you mean motion blur instead
>>108542658Valve has the largest trove of PC game data in the world, so I'm sure it will be fine.
Sounds like something they couldve done years ago
Is there a way to launch my games without launching Steam? I'm on a fairly low end PC and playing while Steamwebhelper on the background consumes 400 megs of RAM even in offline mode.
>>108546629cracked game
>>108541453yeah. it's hard for 45IQ indians to fathom but Valve is the only good company, likely to change when gaben dies.
>>108542658>they work great on my machine.I see this shit posted all the time by people claiming their six year old card "runs fine" but they'll never post a fucking video or even a screenshot of their fps. "Running fine" is subjective and some retards think anything over 20 fps is great.
>>108546629Only buy games that are DRM free. There's actually quite a few that never need steam besides downloading the game.
>>108546845>Valve is the only good companygood goy
>>108540722
>>108544229steam is their only source of endorphins. junkies, essentially
>>108547251my stinky butt is for your consumption
>>108540713it says right there, it averages users with similar hardware to yours
>>108540713Steam has a FPS/Data Overlay options, they can just take the FPS average from that and add it up on X specs, wow.
>>108546283I'd just play on PS5 if I had to do that.
>>108540713Are you fucking retarded? it's already there on your fucking screenshot moron. This is why so threads should be a bannable offense, even a nigger is smarter than you OP.
