/v/ - Video Games

But OP, PS3 has no games

>>16875464

>>16875464

Those are still relatively advanced CPUs, I think (more or less, PowerPC based)…while PS3's Cell is particularly complex.

>>16875502

I'll never consider anything "retro" simply because of age. It's a whole shitty experience that qualifies something as retro or not.

>>16875464

RPCS3 isn't fully functioning, but it works really well.

Give it a try, anon. Seriously.

>>16875502

15 years faggot, I personally consider that the cut-off point.

>>16875493

>Then what the fuck have I been emulating Nier on?

Exactly. RPCS3 isn't a PS3 emulator, it's a Nier emulator.

>>16875512

Then you must think 15-year-old teenagers are retro by that comparison. 25 years is better because it's over past a generation and a quarter of a century old.

>>16875645

demon souls works if you have a good enough cpu

Bitch fuck your 360ps3 everything on those is on PC anyway

Monkey hero

Ps1

Literally never emulatable

I guess "retro" is whatever people remember as a kid.

In which case, I'm an old fart. PS3 was yesterday.

>>16875659

>>16875695

I was half trolling, but not entirely. Aside from Nier, I also got Dragon's Crown and Wangan Midnight running perfectly fine on RPCS3 (though Wangan Midnight is known to crash after every 2 or 3 races, last build I've tested).

I haven't been following RPCS3 closely or testing new builds for about a year, but in the period from 2018-2019 there were tons of regressions, new bugs and incompatibilities introduced. For an emulator to become consistently worse with every new build for such a long period, it's something I have never seen before, and that's the main reason I've lost interest in RPCS3. I wonder if the devs got their shit together in the last 6 months or so but I'm not counting on it.

It goes up and down, depending on a mix of how underpowered consoles are, and how autistic a consoles fanbois are. Early '90s first departures from hardware dependence:

>software-only emulators for different ISAs, good 1st/2nd-gen console emulation, almost-playable 16-bit/3rd-gen console emulators on top-end PCs

Late '90s:

>good 16-bit emulation & some 4th-gen console emulation, barely playable 3D/5th-gen console emulation

Early '00s, apex of emulation IMHO:

>old 2D emulators perfected, first cycle-accurate emulators, all oddball 2D consoles and PC platforms/OSs get playable emulation, 3D consoles (N64/PSX) running new games better than native, some emulators (GBA) 0day actual hardware off leaked specs

Late '00s, nadir of emulation:

>n64 emulation stalls, 6th-gen set-top emus unplayable to nonexistent, even DS/PSP emulation terrible, only bright spot is rise of hypervisors for PC targets

Early '10s false start:

>bunch of unplayable emulators mature, mostly due to Nintendo casualbait hardware (GCN/Wii & DS, to a lesser degree PS2 & PSP), but bunch of extremely promising emulators (360, PS3, Vita, plus seeming breakthroughs in N64 & OG XBox) soon stall

Late '10s rebirth:

>everything, set-top, handheld, and arcade from 7th-gen down (plus 100% of Nintoddler platforms) are emulated somewhere from playable to perfect, with the (deeply ironic considering Xenia's role in breaking assumptions against 7th-gen) exception of 360 which is only mildly playable. Only major black eye remaining is continued inability to emulate 8th-gen consoles

'20s:

>(you) ARE HERE

Only real question is whether emulator devs are going to make hypervisors for 8th-gen console emulation or be faggots like OG XBox emu devs and sit on it for a zillion years. And, I guess, whether the walled garden of mobile is going to be toppled with better console/PC emulators for Android, plus maybe iOS emulation for Android.

On a more minor note, I guess there's tinier quibbles, like the non-bestseller N64 games literally nobody has ever heard of emulating better, certain badly written ( (PCSX2) or abusively developed (Cemu, Citra) emulators get mature competition, and that RetroArch's WIMP GUI becomes capable of completely displacing the XMB UI abomination.

>>16875695

Does it still need to cache shaders? Its been a while since I tried it out.

>>16875698

Duckstation runs it

>>16875732

You are supposed to report regressions retard. Do you expect the devs to read your mind to find out if your game has regressed?

Now one thing I kinda suspected but was actually confirmed recently and really makes me fucking mad is that the one reason why we don't have a very good XBOX emulator right now is that the dumb fucking cunt that is called JayFoxRox has flexed his autism all over the XBOX scene and made it almost impossible to have any amount of decent progress It's sad the guy does know his shit but he literally cannot keep his shit together and has to sperg everywhere.

So now after enduring this fuckery for years the one guy actually doing shit and not arguing about capitalization in commits has made his own project called XEMU, and considering the few vids he uploaded showcasing his private progress on XQEMU

>>16875732

>I haven't been following RPCS3 closely or testing new builds for about a year, but in the period from 2018-2019 there were tons of regressions

It's almost like it's a very complex, experimental and WIP project with two fulltime devs

>>16876175

>Duckstation runs it

Without sound for now in the latest version, still good progress considering that's pretty much the only PS1 game you cannot emulate in good conditions from what I know.

>>16876197

It's wild, you can't even play monkey hero on the ps2

>>16876201

The PS2 BC is really damn good, especially for something only a few guys worked on without anywhere near the current level of knowledge on the machine, granted SONY's internal docs do help but are probably not perfect like most other internal docs.

The weakest link in it is actually the CD drive controller of all things, it doesn't support every quirk the PS1 had and that's why most broken games on PS2 BC have very sketchy CD drive interaction, and why FDS breaks a lot more games.

>>16875502

Because of its prolonged age I won't consider 7yh retro for another 5 years. Only now are people finally admitting that 6th gen is retro

.>>16875506

Return to >>>/cuckchan/ /vr/

>>16875671

>25 years is better because it's over past a generation and a quarter of a century old.

In that case the 5th gen is only now retro and the Genesis wasn't retro until 2013. That nigger logic.

>>16875896

>old 2D emulators perfected, first cycle-accurate emulators

Not really. ZSNES, SNES 9X, Gens, and Kega were far from cycle accurate. Even NES emulators had minor inaccuracies. It wasn't until the late 00s SNES and Genesis emulation was refined.

>>16876266

>Not really. ZSNES, SNES 9X, Gens, and Kega were far from cycle accurate. Even NES emulators had minor inaccuracies.

Note the comma separating those two statements. The emulators you listed aside from ZSNES, which was the sort of "bad" emulator I was distinguishing the newer emulators from became very good even then. Meanwhile, the first cycle-accurate emulators such as BSNES were created.

Been playing with my friends using screenshare app on a ps3 emulator. Works pretty well. It's only a matter of adding more first-party AAA titles that work now.

Could someone give me the skinny on why emulation is so difficult? A modern PC is leagues beyond the PS3/Xbox360 in terms of hardware. Is it effectively encryption?

>>16876344

cpus are not equiped to synchronize threads at the rate hardware modules communicate.

the number of context switches required to properly synchronize various components (cpu, gpu, etc.) is higher than any single/multicore system can handle.

context switching frequently (>mhz rate) make all the features modern cpus have pointless.

early emulators would run each component for a arbitrary amount of time taking full advantage of CPU performance (cache, branch prediction, too many to list), at the cost of units of the system getting information at the wrong time causing bugs.

you can emulate individual components relatively fast in isolation, which is part of why HLE works.

Even the simplest games like super mario bros. require precise timing to run at all, or specific hacks. (read the "sprite 0" interrupt on NESDEV)

>>16876344

Emulation is slow because instruction set are different from modern CPUs and if a CPU doesn't support an instruction in hardware you need to do the same thing in software which is slow as fuck

Also tight sync is hard because high latency between everything in a PC but usually low or near nonexistent in console, syncing several emulated chips tightly means single thread and a massive amount of overhead

Then there's architecture difference like GPU being able to access RAM / CPU able to access VRAM being very slow on modern hardware but no big deal on many consoles.

Then there's also behavior that are completely aliento modern PC like non compliant floating point handling, cache fuckery, chips being able to fudge what another chip is working with with no restriction.

Tl;dr you're asking your PC to do something it's bad at doing because it's generalized hardware whereas the console is made to do it.

PS4 is hard because blackbox APIs even if the hardware is similar (and shared memory between CPU and GPU + not really x86 but sorta x86 doesn't help)

Xbone is hard because low interest with actually decent security

>>16876368

>Xbone is hard because low interest with actually decent security

Xbone is hard specifically because most, if not all of it's games are on PC(even if it is Windows 10.)

>>16876344

>>16876368

Can't find it but one of the good example is from dolphin where they showed that on the console the game tries to run one single small instruction and Dolphin has to basically turn that into 20+ line of code to do the same.

>>16876376

with instruction level parallelism having slightly bloated dynarec code isn't as bad as it would seem. (still fucked the cache I guess).

one small instruction can take countless cycles for a cpu to execute so the translated form is really just a more verbose expression of the semantics.

on x86 you can translate any "CISC" instruction into equivalent "RISC" instructions and have it perform the same (despite more bytes of code).

>>16875896

You forgot one small part of early '10s

>Every emulator needs to be ported on Android, who cares if it barely works, we HAVE to do it NOW

Though it did mean that I could play Ace Attorney with a touch screen.

>>16876524

Well, mentioning that would be getting into the weeds of non-IBM clone PCs and non-MS OSs. For instance, I was a Mac user from the '80s until the Intel transition, and a lot of cutting edge emulators were specific to certain platforms like Mac, Amiga, Towns, MSX, etc. (plus high-end workstation UNIXes), before being ported to wintel, both because wintel was just slow garbage back then, and because some of them had more similar ISAs (less overhead) to non-x86 emulation targets. Then of course there were different OSs even on the same hardware, such as the lengthy struggle of Linux/BSD groupies to get stuff ported.

The Android/iOS ports were a part of this, since they were initially themselves ports from earlier StrongARM OSs like PocketPC, Symbian, & Blackberry, though Android/iOS hardware represented something new in that most earlier ARM hardware lacked floating point or 3D acceleration.

>>16876344

Other answers are quite true, especially for older games, but for some of 6th-gen and much of 7th/8th-gen, console games are written a lot more like PC games, in that instead of games being custom coded, there is a lot of shared code. This allows what is called "High-Level Emulation" (HLE) and special lightweight emulators called "hypervisors".

Aside from being tremendously faster, such techniques are also crucial for making games play better than on original hardware (higher resolution, framerate, textures, etc.).

This is why emulating games between Windows/Linux/Mac has historically been so much easier, because instead of optimizing for each entire games every time, its reused code (game engine, middleware, OS-level libraries, hardware driver APIs) is ported once to native code.

Due to the massive amount of idiosyncratic code in software, as opposed to the far smaller (and more standardized) complexity embodied by hardware, this means the biggest barrier to emulation today is the need for good documentation (whether stolen and "cleanroomed", or simply bruteforce reverse engineered).

>>16875896

>RetroArch's WIMP GUI becomes capable of completely displacing the XMB UI abomination

Just use the rgui menus

>>16876704

Ozone is also pretty good from what little time I've used I've used it for

The real question is why the XMB is the default

>>16876712

Because it's the one that caused the least backlash.

I just use RGUI, fuck "nice" looking stuff, I prefer functional and fast.

**Still sucks ZARCH never went anywhere, it's the PPSSPP lookalike with an

option for ZSNES snow**

>>16876704

>>16876712

rgui & Ozone are both XMB abominations just as much as "xmb". The entire thing of "gamepad-friendly" menus or badly documented hidden .cfg files as the only way to do things on a PC emulator is fucking insane.

I want RetroArch to function the same way as SNES9x, Dolphin, or any sane MAME frontend clear back to the '90s. Mouse, keyboard, buttons, checkboxes, text fields, popup menus, fucking done.

It's not rocket science, but sadly the Qt WIMP GUI is still very incomplete.

>>16876712

>The real question is why the XMB is the default

It no longer is: https://www.libretro.com/index.php/retroarch-ozone-becomes-the-default-menu-ui-plus-touchscreen-and-scaling-updates/

As for why XMB was ever the default, it's because it was the first relatively modern-looking alternative to RGUI, and it received a warm reception when it was first implemented.

>>16876731

RGUI can look pretty nice now that it's customizable.

>>16876743

I agree that it really didn't need to be much more complicated than something like Dolphin's UI. It probably would've been both easier to make and way less bloat than how it is now.

I like retroarch, but its useless for systems like PC/DOS, Amiga, Atari ST.

I can recommend installing RetroPie to get around configuration hell (doesn't actually need an SJW "educational" broadcom blobbed SBC aka. raspberry pi).

I'm quite fond of retro-acheivements, it makes me feel less lonely replaying old games.

>>16875464

if you think that's bad try OG xbox emulation

>>16876801

>I can recommend installing RetroPie to get around configuration hell (doesn't actually need an SJW "educational" broadcom blobbed SBC aka. raspberry pi).

Related question: I'm an nVidia user on Linux (no buli plz), and I'm faced with a dichotomy between nVidia's proprietary drivers (good performance) and the open source "nouveau" drivers (supports KMS/DRM needed for some of LibRetro's coolest features).

Is there a way of installing both drivers, and switching between them without a reboot?

>>16876801

>I'm quite fond of retro-acheivements, it makes me feel less lonely replaying old games.

Unironically fucking kill yourself.

>>16876825

If you also happen to have an intel IGP use that for DRM/KMS.

>>16875464

>Still no fully functioning ps3 emulator

Do you live in a cave or something? you silly troglodyte.

>>16875695

Works fine on my decade old cpu. All you need is a quad with hyperthreading.

But there is such thing as wii emulation though. Play Fishing Master World Tour for one of very very few fishing games that is actually good in this niche genre.

>>16875464

>>16875493

>>16875512

Dragon's Crown works fine on RPCS3 for me. There's wasn't all that much I wanted to play on it since "PS3 has no games" is fairly accurate.

>>16876175

>You are supposed to report regressions retard. Do you expect the devs to read your mind to find out if your game has regressed?

Almost all regressions I come across are already reported by the time I check. Not like reporting makes any fucking difference when most reported regressions remain forever unfixed and new regressions keep piling up with every new build.

Look, I know this sort of shit happens in every WIP HLE emulator but not to the degree and timespan I've been seeing with RPCS3. I'm still using a build from almost 2 years ago because it works better than anything newer. This emulator seems to have gone completely FUBAR and I'd say it's time for the devs to stop for a minute and rethink their strategy or something.

>>16875502

>clover

>>16876266

>In that case the 5th gen is only now retro and the Genesis wasn't retro until 2013. That nigger logic.

actually that kind of makes sense

consider the indie game scene; so many soyfags trying and failing to capture the charm of retro by targeting (and often missing) NES graphics. Only in the last 5 years or so have we seen an upswing in games targeting the SNES and Genesis's capabilities, and just in the last year or two have we seen lowpoly (and the most common misunderstanding of lowpoly at that) emerge among the soy crowd

>>16876344

it's like this:

no matter how fast the CPU, it's roughly built the same

but when you look at a different CPU architecture, you're looking at something that's built different

no matter how good your CPU, if it's trying to run something that's built for different it's going to need an interpreter

sometimes that interpreter is bullshit hard to code

>>16878281

Just get the console if you care that much, OG XBOX aren't even that bad to deal with.

You can thank JayFoxRox for the lack of recent major progress in XBOX emulation.

>>16878281

I'm sure I've seen some OG XBox titles like JSRF emulating pretty much perfect recently, so the situation isn't hopeless. That said, as of 5 months ago at least, Breakdown does not appear to be playable:

https://github.com/Cxbx-Reloaded/game-compatibility/issues/115

>>16878315

It's sadly not super interesting.

Basically JayFoxRox is a turboautist that wants thing done his way or not done at all and also has a tendency to not want regular users to have the nice things.

He's the guy who bogged down citra for years and vehemently refused any and all attempt to support encrypted dumps for a very long time under fallacious pretenses.

He also prevented CXBXR from doing more work on the LLE side of things because they were "stealing" his opensourced work and he cried about it and about how it was detrimental to the continued existence of XQEMU because it "drove user away from the project" while also ignoring that even just setting up XQEMU (nevermind actually using it in it's current state) even if you do download the needed files i a massive pain in the ass not even mentionning if you wanna do it legally which he will witchhunt you about if you're asking for any support (which you will need) through any official channel of XQEMU granted seeing how M$ acts towards those projects I totally understand being extra careful.

But yeah going back to the main point basically one of the big three contributor to XQEMU, MBogerson, did a lot of very good work into fixing and making XQEMU fast resulting in what you can see here https://www.youtube.com/channel/UCPBmoQF6_fYo0FgMt6pZLMQ/videos

but apparently either there were issues getting this stuff in XQEMU or he stopped trying after having to deal with Jay and his autism for years and is only now making his own fork of the thing which might become public SOON™ that way he can do whatever the fuck he wants without endless debates over details.

tl;dr: MAMEdev syndrome

>>16877228

how about demon souls?

>>16878344

>if you wanna do it legally which he will witchhunt you about if you're asking for any support

The cancer that is tryhard moralfag retrogaming strikes again

Quick reminder there has literally never been a judicial precedent set against downloading someone else's copy of something you already own off the Internet

>>16878704

Technically speaking, you're right just on the nature that rips, would fundamentally be interchangeable from each other (greatest hits/v1.1 excluded) the issue is more that a judge might be inclined to feel otherwise.

>>16878704

But there are precedents of projects getting shutdown over supporting piracy and there is also the whole atlus vs RPCS3 thing where they technically cannot name / link / use screenshot of Persona 5 not on any legal basis but simply because they cannot defend from atlus if they go legal even with no basis

Most emudevs do that because it provide a degree of plausible deniability / security from bullshit like what atlus pulled, now that's not to say there's no massive moralfaggy corporate whore but many don't care and most likely do it themselves just in private.

Regardless of how you feel the vast majority of emudev out there won't question you if you're not a retard about stuff you download, meaning that if you don't admit to it and your dump doesn't look suspicious they won't care, it's not hard to do.

Specifically regarding XQEMU though it's kinda bullshit, basicallymost people can't realistically use it legally, because it requires an MCP (basically the mobo's chipset) firmware dump (among other thing but the rest is easy to get), you can't do it in software afaict and I'm fairly sure the process involves complex soldering or desoldering the thing entirely, who the fuck is actually gonna or even be able to do that outside of an extreme minority.

>>16878770

>there are precedents of projects getting shutdown over supporting piracy

Sauce? Every instance I can remember was a "strongly worded letter">"oh noez">"we are grateful for your support over the years" voluntary cuckout that never, ever reached trial.

>basicallymost people can't realistically use it legally

That's bullshit. If you own/trashed XBox hardware, you legally own XBox firmware, simple as.

>>16878787

>Sauce?

Mostly playstore fodder getting banned for having ROM downloaders or ROM sites links included, there is also a few web based ones that got C&Ded for similar reasons, pretty much the only ones that never caught much flak for doing it are iaMAME ad that one netplay fork of MAME I forget the name of with P2P ROM sharing.

>you legally own XBox firmware

Yes, but you only own the one inside the XBOX strictly legally speaking, so you technically have to extract it yourself from your hardware, in practice it's almost impossible to prove you did or didn't extract it yourself (since it's not unique per console) and since there's technically no case involving that and that have had a verdict that I know of in the US no one knows what would actually happen.

>>16878802

>ROM downloaders or ROM sites links included

Yeah, that's a different kettle of fish from writing an emulator, or finding ROMs to download yourself

>but you only own the one inside the XBOX strictly legally speaking

No court nor law has ever affirmed this nonsense

>>16877596

It'll be a cold day in hell before I let some literal cucks dictate what is and isn't retro. 6th gen is the most recent retro generation, regardless of what boomers or soycucks say.

>>16878808

>No court nor law has ever affirmed this nonsense

Like I said no court but legally I'm fairly sure there no exemption from any law for downloading a backup of your copy whereas there is one for making you own backup, and then there's a fuckton of legal bullshit / stupid technicalities that could be grafted on a case if such a thing ever went to court.

>>16878770

Why does fatlus even care? They've made so much money off P5 that they're probably wasting more than they'd gain by going after rom sites and emulators.

>>16877545

>Almost all regressions I come across are already reported by the time I check

>I'm still using a build from almost 2 years ago because it works better than anything newer

Name a few issues then

The oldest regressions are either just untested or have incomplete information.

speaking of emulators, Shadow of Rome on PCSX2 runs super slow for me, as does Sly 3. I have AMD gpu (580 Rx) and cpu (Ryzen 5 1600), could that be it?

>>16893357

>mummy, my amd shit is shit at emulation, what is wrong with it?

>>16875464

no, not even PS2 emulators work properly still

>>16875464

>Will emulation of old games ever be perfected?

yes. ask byuu.

>>16875698

try it on Xebra. If that doesn't work then it truly is unemulatable