Instalar Steam
iniciar sesión
|
idioma
简体中文 (Chino simplificado)
繁體中文 (Chino tradicional)
日本語 (Japonés)
한국어 (Coreano)
ไทย (Tailandés)
български (Búlgaro)
Čeština (Checo)
Dansk (Danés)
Deutsch (Alemán)
English (Inglés)
Español - España
Ελληνικά (Griego)
Français (Francés)
Italiano
Bahasa Indonesia (indonesio)
Magyar (Húngaro)
Nederlands (Holandés)
Norsk (Noruego)
Polski (Polaco)
Português (Portugués de Portugal)
Português - Brasil (Portugués - Brasil)
Română (Rumano)
Русский (Ruso)
Suomi (Finés)
Svenska (Sueco)
Türkçe (Turco)
Tiếng Việt (Vietnamita)
Українська (Ucraniano)
Informar de un error de traducción
Until someone optimizes per app like is being done in linux, or gives a plug and play OpenCL interpolation option since they replied (LinusTech Twitter), the actual group, to my accusations of its non existence with proof it does, the advantage is going to be extremely limited and extremely hit and miss. Think old game engines or titles the physics are fussy girls or ones the CPU cant even maintain a 33.33 frame update rate without dropping.
You can skip steps completely which is a speed up, but the real one is the ability to place mickey mouse graphics API overhead amounts on another thread that is in no way really fully utilizing it, and I dont think this works at all if the game is coded to use one core. Therefore DX10 and under is the use case but that falls off pretty quickly too. Now it can be an improvement. StarCraft II the lows with CPU bottleneck (Draw calls are also a factor) Vulkan in my testing is faster. Is it really by a lot? ♥♥♥♥ no. Most people arent going to be able to tell the difference once its fully cached the shaders which is another at least rare annoyance. Also who is playing 4 vs 4 with randos except for me. Good benchmark if armies mass late game. Replays by player camera if you turn off the non deterministic physics is a good way to test lows.
Right now the potential of the technology hasnt been exploited, and the Steam Deck commands with its latency situation meets linux performance overhead under gaming conditions, this is taken more seriously as a mass optimization vector. It really needs to extend beyond Linux since everyone knows windows is the superior gaming platform. Even Valve. Even if investing in Linux is the right choice, just not exclusively.
I am not a programmer, but thats the end game. To me not including another layer of cache, like AMDs infinity, halving the L3 on Zen 2 which already has latency issues architecturally, and early bins of low power DDR5 is going to invite older less optimized software to be undesirable. Which honestly should also be the machines domain of play since its GPU in practice is only going to really beat the GTX 750 Ti im making presets specifically for due to its popularity out in the real world. So in a way im a developer for the Steam Deck, and I know its limitations are a great opportunity to take the software side to the next level. Its CPU might be iffy in some cases, but you can also sort of cheat a little bit if you have insane compute shader performance, which it does for what it is, and the right software to make that an advantage.
In the end, I will probably make electron preservation sensitive presets since the standards I have demand multiple radii within limited screenspace only dynamics, another thing Valve could easily solve.
Examples:
1936x1089 off borders unseen by mankinds eye
1920x1080 actual
1296x809
1280x800
...and just plug the overhead pick up with a clever use of FSR to have your cake and eat it too. Its a dirty hack for beyond screenspace only resolve which multi radii and human tuning would have further grace.
The reason im putting so much work into this project is I know it will eventually attract Morpheus. Right now I need what should be simple A.I. texture upscale PC Master Race Scientists to chip in. This could be implemented in a more widespread way. Even DG Voodoo can pass along its capabilities into DX11. This gives hints that a programmer could make texture upscaling through graphics API post OG compression intercepts, that works in a massive library across Steams entire 3D ecosystem.
Honestly modding should be given more investment from Valve itself very directly from an objective area of rapid self improvement like skynet angle of parallel universe realized. Im not doing this to get rich, im doing it because probabilistically it being done right in a consistent scientific manor, is hopelessly unlikely. I havent even gotten to the noob level of optimization which isnt programming tier by any stretch but possibly the absolute hard limits on what is possible from the outside. Just making the graphics component, and doing that right and tuning it for hardware people actually own and getting it to work with as many modern software buffs as possible is insanely time consuming. Basically im not looking for a hand out im looking for parallelism. All my work is open source or its entirely pointless. Far too niche and abstract for the average person to care. Essentially what I am doing is envisioning objective patches for ancient software in a broad spectrum way. Several games out of the box from the Steam store are extremely broken=Enslaved (Performance bug default renderer windows), Max Payne (Steam Deck CPU incompat), Final Fantasy XV (Performance is a dumpster fire).
If the Special K character who doesnt like reshade compatibility but implies it was at minimum an attempted coexistence setting traps to break his software basically if you actually want functioning effects... if he can improve the latter mentioned games performance wildly, there is really no excuse our God Newell should not personally intervene. As it stands right now I cannot recommend the Steam store version of this game. Get the Origin version. I get Valve has priorities, and they are more important than Final Fantasy XV as a general example of what one might call irrelevant history, but for how long? Eternity? Is it literally up to blips in behavioral genetics to rescue mankind from this blatant mediocrity?
Why am I the pinnacle of youtube retro enhancement without personally building games in new engines using copy and paste? -- and this will be undeniable once I crack BMP to TPF conversion and understanding of the hashtree structure multi asset singularity of name thing. Probably a good reason for that. Im a hard worker and I actually care about the nuance. Most people have the check box of, does it fully populate my 4k screen, or whatever that is. Nasty my performance just cut in half. I of course could do better. Adoption is all about how much penalty incurs. Its why I have optimization as a eventual focus. People are fickle beasts as a rule. With battery involved, its going to have to be a masterpiece of optimization. For that to become reality im going to need passion on high. Amen.
Whats funny is something Daves Garage plotted out before my eyes days ago is being more fully appreciated in real time. Smart guy, thought he was just being funny, but actually both that and literal. If only we could clone an army of his phenotype to upgrade the entire world. Honestly if we had something that was just automatic like microsoft invented for A.I. Texture Upscaling that worked in literally any API that would save me a lot of nightmares of being ingested by the black hole that is Valve time, but where it completely stops if peering into that sort of Medusa. Sounds like a deal sweetener that cannot be ignored. Especially if it applies to emulators. Absolutely no one who is sane is going to want to play the massively inferior histories in the Jupiter that is Steam. The power levels of the Steam Deck will only mature with this transition of fate. It would be a broad spectrum reason to buy ancient video games, which at the end of the day is good for business.
Make it a new rendering solution like when software jumped to blended textures and mipmaps gave way to anistrophic filtering which has now evolved into A.I. model upscaled packs you can download as optional, or just train up your own pack yourself in real time at the cost of some performance in the short term. This would optimize bandwidth and convenience. It would absolutely be worth developing your own implementation using the Open Source models available for A.I. upscaling. Just build around that and make it possible to standardize, including the 4GB patch by automation in a precheck intelligent manor. Also a per app check from a living human being in an archived library sort of way for texture compression to sort of feel out the limits for compatibility reasons would make sense if you had a hard cap on say year 2015 and under.
In the future if I read correctly it sounds like DG Voodoo will be able to enable Truform into any game it hooks with. Not efficiently, but he said it was possible to use human intelligence to flag where you want it to ignore. The OG Glock from Half Life being a great example. And throughout the LOD render. Im thinking of this technology without the emulation component as a universally applicable tool for enhancement protocols overseen and developed only by the most advanced of modders in the community. Thats without question going to be the tool makers. The pinnacle of technologists. Im just sort of a custodial librarian type.
Hear me, oh Lord of PC Master Race Cognitive Processes. Emperor of innovation, unleash your full fury on the enemy. I have detected massive waves incoming from the deepest blackness of space. 60% relativistic speeds and slowing down. The Vangard of advanced peace (Zen 2 Steam Deck confirmed) is requesting back up. Erasing the mediocrity of their first fleets has been very time consuming Lord Newell. They have self replicating technologies, so send back up to be sure. Quantum communications out.
"One of the things that's interesting is that the PC has always had a huge amount of scalability. It was sort of the wild dog that moved into Australia and killed all the local life because it could just adapt. There used to be these dedicated devices, like dedicated word processors."
-God
Amen.
But as you said, there is very little reason to use it outside of Linux gaming, since, well, directX 8-12 all run without problems on windows 10 at least.
Don't know if dx7 is possible too. I think it is, because I've seen someone playing team fortress 2 with that api. But maybe he or she used windows 7 or 8.
But at least on Linux, everything works fine (including the wine part)
Ok and to thise guy with that big chungus long text
I'm not gonna read this. Dear god my eyes when i just look at this
https://imgur.com/a/ysYmeoJ
Top picture = Windows
Bottom picture = Linux
DX12 Layer on-top of Vulkan for Proton is located here:
https://github.com/HansKristian-Work/vkd3d-proton
I really hope intel arch will also have open source drivers such as their current i-gpu's.
Also no problem with my crap-laptop running some intel celeron with hybrid ♥♥♥♥ graphics. (At least good enough for school)
Once you have this mentality, you don't want to miss it.