Silent Hill 2 Remake Performance Benchmark

[GodisanAtheist]

View attachment 367028

- Super Mario Bros is making that 3090ti SWEAT

 

[Vayra86]

That's the same for 99% of ALL modern AAA games, high/medium settings and tweaking will give you 95% of the same visuals as the GPU destroying Epic/Ultra settings, there is a corner of gamers who are snobs and think that they should be able to run every game at 4K/RT/Ultra settings at 200+ FPS, further exasperated by the introduction of ultra high Hz monitors and CS/FN kiddies who think that running 300+ FPS makes them more 1337 than everyone else, obviously never had to deal with trying to hit 30FPS at 360p back in the day or your games not running at all cause you didn't meet the HW req's

There are examples on here as well as everywhere else, complaining their games look shit and unrealistic unless they are running at 4k and 200+ FPS, pure snob/twatery IMO if you need that kind of bullshittery to play games then you should find another hobby, those and pixel peepers analysing every frame, it's a mental illness IMO, play the game and enjoy it or don't, if you look for fault in everything you will find it and also have a sad fucking life

It could be that but games IN FACT become worse from the visuals standpoint. Yes, textures are much higher definition but TAA, Nanite, Lumen etc shenanigans only make games putrid vomit and optical migraine running at abysmal framerates. People like you are enabling this cancer.
Gamers should be much more vocal about rubbish state of modern gaming. With equipment like, say, 4070 it's nonsensical to "run" games with visuals worse than in GTA V at 1080p at 50 FPS. There is no excuse for that. UE5 must be abandoned and UE6 must be a completely new product written from scratch that contains zilch cancer introduced in latest years. But alas.

I think you are both quite correct, honestly. The bottom line really is though that graphics have taken a massive step back. This isnt a matter of diminishing returns; and its not caused by the few percent of pixel peepers either; its pure inefficiency made reality and a lot of games are not really driven by creative talent. The visual gain is there but doesnt make for a better graphical experience. Art style and creativity is where its at; completely unrelated to hardware. You can stack six hundred implementations of upscale and other nonsense on top but turds will still render as turds.

 

[Veseleil]

Just another UE5 garbage.

 

[Trunks0]

Looks like there is a pretty decent Ultra Plus mod with a laundry list of optimizations, tweaks and fixes.

Ultra Plus Fixes and Visual Improvements

An all-in-one fix for stutter, particle trails, fog noise, and reflections with visual improvements. 100% of donations go to Wildlife Society.

www.nexusmods.com

Can't personally test it... the price on this game is bonkers. But maybe it'll help some one out or make for an interesting article if @W1zzard was interested (they have also done allot of these UltraPlus mods for other games, would be cool to read an interview or something)

 

[Godrilla]

No lol.

I guess it would be difficult for the biggest ai company in the world to create a custom algorithm for machine learning for it cloud geforce now 4080s and 3080 that runs a simulation that scans for anomalies via crashes, memory leaks etc etc. It would also be difficult to get user feedback that is built in to its Nvidia app /geforce experience via telemetry to prioritize the optimization of its driver development.
I am not talking about an auto pilot level optimization here. I am talking about a hybrid between ai analytics that allows the human engineering development team to prioritize what to optimize first as the rate limiting step in the pipeline. If it's not a thing I believe it is plausible. Let me know why you believe it's not.

 

[Vya Domus]

Let me know why you believe it's not.

Optimizing drivers in the context of performance means taking shaders and rewriting them or modifying the compiler to emit more efficient code, this isn't the type of stuff that can be automated. I am sure they have software that scrubs through bug reports and telemetry but that's not "optimizing drivers".

 

[Han44]

New game for new rtx 5000

 

[Makaveli]

New game for new rtx 5000

not sure I agree there is nothing visually exceptional about this game.

 

[Han44]

New game for new rtx 5000

not sure I agree there is nothing visually exceptional about this game.

My only point was that a high-end GPU cannot cope without boosters in frame rates without boosters at higher resolutions. Nowadays, you can buy a GPU for a lot of money, but without boosters it's a pity. In the past, cards could function without boosters and games were refined and more optimized.

 

[Makaveli]

New game for new rtx 5000

My only point was that a high-end GPU cannot cope without boosters in frame rates without boosters at higher resolutions. Nowadays, you can buy a GPU for a lot of money, but without boosters it's a pity. In the past, cards could function without boosters and games were refined and more optimized.

totally agreed not a fan of having to use upscaling and FG as a crutch for not optimizing games properly for hardware.

 

[AnotherReader]

Optimizing drivers in the context of performance means taking shaders and rewriting them or modifying the compiler to emit more efficient code, this isn't the type of stuff that can be automated. I am sure they have software that scrubs through bug reports and telemetry but that's not "optimizing drivers".

Compilers are still dumber than experts at optimization; this is a big part of why Itanium failed. Almost all game ready drivers include hand optimized code for the critical shaders of a game. The quote below is from the article linked in the first sentence:

I’m guessing RDNA 3’s dual issue mode will have limited impact. It relies heavily on the compiler to find VOPD possibilities, and compilers are frustratingly stupid at seeing very simple optimizations. For example, the FMA test above uses one variable for two of the inputs, which should make it possible for the compiler to meet dual issue constraints. But obviously, the compiler didn’t make it happen.

 

[Arkz]

People have been saying this for over a decade, and it has never held true.

People have been saying games should be optimised for the RDNA tech in the nearly 4 year old consoles for over 10 years?

 

[GodisanAtheist]

People have been saying games should be optimised for the RDNA tech in the nearly 4 year old consoles for over 10 years?

- PS4 & XBone both also used X86 CPU cores and Radeon based GPUs (GCN 2.0 tho) and games still ran faster on equivalent Nvdia GPUs then too.

AMD fans tend to lean on the "But AMD is in the consoles so everything will be default optimized for AMD hardware" and its basically never born out (or its really sad that even with default optimization NV cards still generally overperform).

 

[THU31]

I assume TXAA is a typo in the game, and not actually NVIDIA's tech from a decade ago?

 

[Melvis]

Because people are interested in it now, and not in the future. Source: Steam number of player charts

People are silly and impatient! and does that source tell you how many stopped playing the game after an hour of playing it realizing it is a sloppy, some times unplayable unfinished mess? Ive learnt the hard way ive been there before so I never buy first day releases etc! and I wait! as lets face facts there are plenty of other games to play in the mean time and then buy it at a later date, most likely cheaper! and alot less bugs and to them/me its still a NEW game.

 

[Vayra86]

- PS4 & XBone both also used X86 CPU cores and Radeon based GPUs (GCN 2.0 tho) and games still ran faster on equivalent Nvdia GPUs then too.

AMD fans tend to lean on the "But AMD is in the consoles so everything will be default optimized for AMD hardware" and its basically never born out (or its really sad that even with default optimization NV cards still generally overperform).

Optimization no, stability and good performance yes. But optimized titles are few and far between anyway , imho there are just varying degrees of crappy now. If you see the look vs the perf cost, we are solid in the land of ridiculousness. Not stagnation, but degradation really, with a bunch of crutches keeping it somewhat playable.

 

[W1zzard]

how many stopped playing the game after an hour of playing it realizing it is a sloppy, some times unplayable unfinished mess?

it does: https://steamdb.info/app/2358720/charts/#1m

 

[Prince Valiant]

Mediocre performance for the visuals on low and the trade only gets worse from there .

I dont understand why people benchmark brand new games when these days there clearly full of bugs and are a unoptimised mess! Give it a few months when these are mostly sorted and the performance numbers will go up for sure as they usually do for both Nvidia and AMD.

They offered it in this state, it's fair game.

 

[Arkz]

- PS4 & XBone both also used X86 CPU cores and Radeon based GPUs (GCN 2.0 tho) and games still ran faster on equivalent Nvdia GPUs then too.

AMD fans tend to lean on the "But AMD is in the consoles so everything will be default optimized for AMD hardware" and its basically never born out (or its really sad that even with default optimization NV cards still generally overperform).

PS4 was about equal in perf to what, a GTX 660? I'd say it did pretty damn good for that. Can't imagine Horizon Zero Dawn or Forbidden West running any better unless you paired it with a much meatier cpu than that baby jaguar chip the Ps4 used.

 

[nguyen]

it does: https://steamdb.info/app/2358720/charts/#1m

I checked the global achievements for Black Myth Wukong and a whopping 40% of players finished the elusive secret ending, with 21mil players total that would be ~8 millions players that have completed the game.

 

[GodisanAtheist]

PS4 was about equal in perf to what, a GTX 660? I'd say it did pretty damn good for that. Can't imagine Horizon Zero Dawn or Forbidden West running any better unless you paired it with a much meatier cpu than that baby jaguar chip the Ps4 used.

-Right but I was specifically speaking to AMD optimization. Who cares what the PS4's power level is, the point is the games that come to PC still tend to perform better on NV hardware despite being built for AMD hardware in the consoles.

Mediocre performance for the visuals on low and the trade only gets worse from there .


They offered it in this state, it's fair game.

-Would be nice if we could get a revisit of some of these games 1 year out from launch. Not everyone buys day of, and a significant number of people wait till they get snag a complete edition for $10.

 

[The Norwegian Drone Pilot]

I predict a patch will come out soon to fix that.

 

[Vayra86]

I predict a patch will come out soon to fix that.

Yeah there is that occasional major stutter going on. But honestly? Whatever the fck ever. The OG versions ran at some order of 25-30 FPS per second, if enemies came close and had to be rendered you had the good ol' massive stutter that was common in early 3D games. This game doesn't really punish you for any of that. Its a slow game anyway. But yeah in the grand scheme of things, with current day tech, its inexcusable.

 

[Prince Valiant]

-Right but I was specifically speaking to AMD optimization. Who cares what the PS4's power level is, the point is the games that come to PC still tend to perform better on NV hardware despite being built for AMD hardware in the consoles.

-Would be nice if we could get a revisit of some of these games 1 year out from launch. Not everyone buys day of, and a significant number of people wait till they get snag a complete edition for $10.

I'd be curious to see such tests. Not sure any outlets would want to run them given the time sink even with top cards (maybe low or bottom end cards too) and letting the reader extrapolate.

 

[GodisanAtheist]

I'd be curious to see such tests. Not sure any outlets would want to run them given the time sink even with top cards (maybe low or bottom end cards too) and letting the reader extrapolate.

-Is love to see a revisit with top/mid/low cards so 3NV/ 3AMD/ 3Intel(maybe).

9 Cards, 1 game. Doesn't sound too rough to do every once in a while and it would bring readers to the yard.

Also do a Top 10 games currently played on Steam review on a regular basis (so long as the games actually churn).