Metal Gear Solid Forum  /  The Future of Emulation for the Leaderboard
  PlywoodPlywood
(edited: )

Hi all,

Ongoing discussion in the MGSR discord has brought into question a handful of aspects of emulator on the leaderboard. I will try to present it here as objectively as possible.

Recently, DuckStation was added to the board as an emulator with the caveat of underclocking the CPU to add in frame lag. This frustrated some runners as not only was the CPU underclock amount dependent on your particular PC, but not every emulator currently allowed on the board has similar lag behavior (Mednafen and Bizhawk have a more steady framerate than DuckStation underclocked or epsxe). Additionally, it represents another setting you have to tinker with before getting into the game, which acts as a barrier to entry.

Thus the point was brought up that DuckStation should be kept to default CPU clock speed, which is fairly similar to the game's performance on digital platforms.

Regardless of this restriction, the conversation of whether or not emulator should be on the same board with official hardware came up. As it stands, all playstation platforms battle it out together, forcing the user to filter to compare based off of similar speeds. This is both fair (the honest truth that whatever is the fastest stands at the top, meaning PSPGO/PSP/PSTV/PSVITA), and unfair (if you are playing on emulator, there is a lower ceiling of best times compared to faster hardware).

As time goes on, the possibility of fast disc speed style emulation is becoming more and more of a reality, with a combination of CPU overclocking, disc caching (in the case of retroarch beetle), and disc speed multipliers. While inaccurate to PSone loads, this would be a great quality of life improvement. However not everyone's PC is up to the task of overclocking the emulator CPU or caching ISO in RAM. This would create a similar hardware arm's race problem to official hardware.

We have also seen some boards in recent times (Dino Crisis 2, Resident Evil 1) require DuckStation ONLY for their emulator board. https://www.speedrun.com/residentevil/thread/5zg8d The concept of requiring only one emulator (or a more narrow set of options) could be done for this game to keep everything on the same level as possible, but would reduce the possible options for the player.

So, there are multiple points for discussion here.

1. Should emulators be kept to default settings as often as possible, even if that means sacrificing 'accuracy' in certain respects? Keep in mind that there are certain settings we'd still need to control for.

2. Should emulators return to being a segregated leaderboard to create a platform for:

a ) as similar to original hardware in terms of loads as possible
b ) aim for fastest load times, essentially disregarding how emulators have been handled for the past 4 years?

3. Should certain emulators be sunset (for example ePSXe, older versions of bizhawk/mednafen) to reduce the complexities of leaderboard management?

Please respond here with your preferred option. If your preferred variation isn't the most popular, we will take that as a vote for the Option in general. Explanation of your reasoning is not required but is requested!

Option 1: Emulator settings become more streamlined but do not get separated from hardware.
Option 1 Alpha: Deprecation/Sunset of ePSXe/etc.
Option 1 Bravo: Keep ePSXe/etc.

Option 2: Emulator settings become more streamlined but DO get seperated from hardware.
Option 2 Alpha: Maintain psone load speeds, but not in game CDROM frame lag.
Option 2 Bravo: Change the board to Fast Disc Speed style emulation

This poll will be up for a week, after which we will adjust the leaderboard accordingly. Regardless of what happens, my testing of RetroArch Beetle shows it to be very good, as long as certain settings are used. It could be key in the circumstances of Option2B being the most popular.

deserteagle417deserteagle417, Hau5testHau5test and 3 others like this. 
  edboiedboi

Option 2 B. I think separating all platforms would be a mess and I don't want this board to become MGS3 lol. So I think adjusting Emulator settings to make running Emu better and as fast as PSTV/Fast Disc Speed while still unseparated would be the best.

PlywoodPlywood likes this. 
  PlywoodPlywood
(edited: )

Worthwhile mentioning as well for further context about Fast Loads in MGS1 console:

PS2 Fast Disc Speed's stability entirely depends on the PS2. Some PS2s commonly crash using FDS (elevators, certain loadzones, etc.). Plus, PS2's FDS is the slowest of the platforms that allow it. While PSTV FDS is more stable, the cost barrier, increased input latency and chance of input drops on PSTV annoy people. Anyone familiar with PC emulation knows that emulators have a certain amount of input lag due to emulation process itself, but from general consensus PSTV latency is more variable.

FDS Emulator could introduce a certain level of complexity to setup for the runner, as a disclaimer.

edboiedboi likes this. 
  dlimes13dlimes13

My vote is for option 2B. I will speak more for DuckStation. Default settings with CPU set to 333% and Quad Disc speed created near PSTV like loads without sacrificing stability. This was tested on a high-end PC (i9-9900K) by me and a relatively low-end PC (i5-2500) by JMC and both had similar results. If we do go with option 2B it will create a more level playing field for those who don't want to enter the console hardware arms race that is PSTV, but wants to play the console version.

I think keeping the settings as close to default with the minimal tweaks for DuckStation as suggested above makes it easier to understand for everybody.

We should also sunset most other emulators. Maybe DuckStation and RetroArch Beetle would be the only two approved emulators in this circumstance. This remains to be seen.

bastianfrusciantebastianfrusciante, The-Wavy-CommanderThe-Wavy-Commander and 2 others like this. 
  ApacheSmashApacheSmash

Option 2B - Aka Operation Fuck the PSTV.

The-Wavy-CommanderThe-Wavy-Commander, PlywoodPlywood and edboiedboi like this. 
  Lime_MeringueLime_Meringue

The post perhaps doesn't have the wording I'd use however I'm all for emulating a faster load within specified parameters, such as emulating FDS. To be honest, even with emulating FDS I wouldn't mind keeping them on the same boards and just adding a either two emulator categories (SDS and FDS) or even adding a separate field for FDS, that way we also know which PS2, PSTV, PSP, PSVita runs are done on FDS too. This would still leave PSTV as the kingpin though which may not be ideal when we are trying to reduce the barrier the entry as well as the barrier to competition.

Considering the ease of use with setting up Duckstation, I'd say option 2b would be my preferred option.

PlywoodPlywood likes this. 
  ShardOfKingdomsShardOfKingdoms

2b seems to me like it would be the best for accessibility, and less annoying

PlywoodPlywood likes this. 
  [user deleted]

For me
Option 2A
I don’t think emulation, by its very nature should run faster than the hardware it’s.. well emulating. That’s literally running in direct contrary to their point. Separating emulator makes a more fair playing field, you don’t have to compete against PSTV, basically. So as far as I’m concerned there’s no need to run them at settings which decrease load speeds dramatically or parrot the PSTV.

I’d also like to say, while I’m not a hugely active MGS1 runner, I’ve been around MGSR for a long time & scarcely (if ever) does a decision, especially one of this gravity please all sides.

I’m aware a lot of folks don’t really like needing a PSTV to be competitive, but again, separating emulator (aka 2017/18 MGS1 boards) makes that point moot. You’re comparing emu to emu and official hardware to official hardware. Allowing load time increases does not do anything more than improve the QOL of the run on emulator, but this alone is not in my opinion a good reason enough.

Thank you for reading!

PlywoodPlywood and hoonsbyhoonsby like this. 
  Solidstrife90Solidstrife90

I think that as much as we have tried to advocate for the other options, 2B seems like the only viable option going forward.

Option 1 us always going to have more differences within the same board due to the range of differences in hardware. And let's face it, MGS1 isn't one of those games that's really hard to get hold of physically compared to other games (from a supply point of view, not monetarily).

2a is tricky to get exactly the same as console, probably takes more work to verify, and if you're going down that route, may as well be added to the console board. That opens up the board to more variables as discussed. So 2B, just separate it and let emu be emu

PlywoodPlywood likes this. 
  iLL_PazzoiLL_Pazzo

If it's possible to get emulation and PSTV really close in terms of speed, I would say 2B : getting emulators and console runs on the same leaderboard, because there would be no point seperating them. But it has to be really close, like max 1 minute of difference not much.

If it's too complicated or impossible for whatever reason, I would go for option 2 : get hardware and emulator separated. The only concern I have is getting everyone on the same leaderboard on the same game speed, to make the console MGS1 runs balanced and fair. As I already said, a leaderboard should reflect as much as possible the skill of the runners imo. Just like it is right now on the PC leaderboard. And I think if everyone is competing on the same material, it would improve the level of the leaderboard by much.

Again if it's not possible to get emulators and hardwares on a nearly same speed, I would go for seperate them. Because not everybody can/wants to spend a lot of money on a PSTV (which btw is stopping its game selling really soon).

bastianfrusciantebastianfrusciante, PlywoodPlywood and 2 others like this. 
  SolidSocomSolidSocom

I'd go with what pazzo said if there isnt much difference same board if there is separate them

so 2b if its close if not just 2

 
  JMCJMC
(edited: )

I'd vote 2B, aim for fastest load times, accuracy clearly isn't what the official solutions went for so why should we lock our 3rd party options to ps1 accuracy.

In an ideal world, I'd vote for a complete split official digital / official disc / duckstation.

That won't happen here so 2B is the only sensible solution.

bastianfrusciantebastianfrusciante and PlywoodPlywood like this. 
  Hau5testHau5test

I'm a fan of 2B as well, emulation is its own beast and I'd like to treat it the same way.

PlywoodPlywood likes this. 
  PlywoodPlywood
(edited: )

I vote for 2a.

The purpose of psone emulation is to do just that, emulate the original psone. The fact that official solutions do not match original psone loads with Fast Disc Speed (a setting provided by Sony) is irrelevant to me because it is an officially released Sony TM solution. I don't care how bad the emulation is by default (e.g. PSClassic), it should be on the board just as much as any other version. An unofficial solution by comparison that doesn't pass a certain bar is not worth our time (e.g. Bleemcast).

This is a position based off of my personal biases: official solutions > unofficial solutions. You may consider it an artificial barrier, but it keeps competition simple, standard, and expected in terms of emulator.

Separating out emulator fulfills the desires of some to avoid comparisons to console and have their fenced in competition. As far as the board is concerned, you have achieved a very high quality of life because the segregation makes your runs relatively faster and elevates you over an actual psone! The addition of FDS emulation puts aside all the runs done on emulator so far and 'resets' the board. If we are fencing in the competition, why put aside the years of emulator runs? Compete against those!

Since it was brought up here, I will explain why the Crash Team Racing orientation of EMU/DISC/PSN is not a great fit. PS2 Fast Disc Speed (the fastest way to play DISC) is known to be unstable for MGS1 depending on the PS2 model and some other unknown factors. Creating a competition centered around an inconsistent solution sounds pretty bad in the long term. It also means PS3 Digital is stuck with PSN, which from a fairness perspective is pretty silly. So either you make separate boards for every console variation, or just accept what is fast and what is slow.

 
  NickRPGreenNickRPGreen

I agree that 2B seems the most sensible. Attempting to match PSTV speeds whilst keeping emulation and console together could result in some difficult verification issues. I'm not convinced that the stability between different emulators/machines has provided fully satisfactory results that speeds will be consistent, though making the call to retire some legacy emulators would certainly reduce that risk. I don't see an issue with limiting the emulator list further, if a runner is interested enough in competing, it's not difficult to download a specific emulator to run it on.

By making the separation, it gives us the wiggle room to alter emulation settings and speeds without compromising the integrity of official hardware runs in the future. I also think the community in this category is sizeable enough to warrant the split without making any category look barren.

bastianfrusciantebastianfrusciante and PlywoodPlywood like this. 
  DeveuhDeveuh

tl;dr: 2.B

Hello,

I'm relatively new to MGS Speedrunning (and in Speedrunning in fact) so maybe I don't have all the knowledge to take the right decision but here my point of view:

For me speedrunning is about speedrunners, not hardware: speedrunners have to train to improve their time, not by constantly looking for faster hardware / settings. In an ideal world we could have just one hardware or multiple hardware / emulators but always with the same performances (loading time / framerate etc...). As I understand here this is not (and can't be) the case.

This being said, I think option 2 is more reasonnable here...

Regarding 2.a or 2.b, my point is that we don't care if emulators are faster than consoles or must mimic the consoles loading time (I don't see console as the superior category that must be ahead emulators). So option 2.b sounds better for me, but with two warnings :
- All emulators must have same performances (either by allowing only one emulator or be looking for counterpart settings accross emulators). I think speedrunning must not be an hardware war.
- Selected emulators (and settings) must run on relatively low-end confs as MGSR must be accessible 🙂

Last thing, if we select one or many emulators, we have to take care about updates for these projects that could break the leaderboard.

I hope my post is understamble, english is not my native tongue 🙂

Thanks,

bastianfrusciantebastianfrusciante and PlywoodPlywood like this. 
  deserteagle417deserteagle417

I don't really have a horse in this race since I don't run emulator so I don't really have an opinion on 2a vs 2b, but I did want to clarify one thing for RE1 and Duckstation. RE1 didn't have a workable PC version for a long time, so emu was the main category for competition. What I did there was actually make a completely separate emulator category just for Duckstation. You can still use the previously approved emulators in the old category, but Duckstation would only compete with itself in the new one. This was done to preserve the history of the board and also give people a real competitive board for a modern emulator.

PlywoodPlywood likes this. 
  hoonsbyhoonsby

Gonna go with 2A.

PlywoodPlywood likes this. 
  MetalGlennSolidMetalGlennSolid

I must agree that 2B seems like the better option than arms racing ourselves (which is what keeps me away from console personally from a competitive standpoint); I do not think under any circumstances that emulation should be faster than the load times of the fastest existing hardware; whether that be PSTV (which will now cost an arm or a leg currently *at least for me at $250 on eBay) or whatever else in the future. The fact this can be achieved by an i5-2000 series CPU and still remain competitive should the runner desire to compete for top times is very nice; meaning this emulator can almost match the fastest existing hardware on a potato is something that is very encouraging in my eyes.

ApacheSmashApacheSmash and bastianfrusciantebastianfrusciante like this.