Software Support > GroovyMAME
Games with ISSUES: report here!
Recapnation:
I recall someone looking into CP-S speed issues recently, though? Don't know how it ended or if it did.
--- Quote ---I wasn't serious regarding the bounties, chances mamedevs would consider taking paid jobs like that are probably extremely low, for the PITA development reasons, but also because in the event that would actually happen, they'd expect way more money than we're ready to pay.
--- End quote ---
You should check Retro Arch's Patreon to get an idea on the numbers. But it's just a matter of they asking for once. There's interest. Garou (be it wait states or whatever the cause), Hyper Fighting, Progear no Arashi, CV-1000... Too many significant (and popular) games now not to get many people aboard, even if it' just for historical reasons. But they need to explain it properly. A hidden webpage in a site called "MAME Testers" where the subjects aren't properly addressed, reports are kept open when they shouldn't, the search system is nuts for non-testers, language gets too technical to explain elementary bits related to performance... isn't obviously the best place to make this stuff visible.
CP-S issue, for instance:
https://mametesters.org/view.php?id=408
Not to mention the uncountable places where these subjects are discussed by the people involved -- impossible to keep track.
The discussion belongs elsewhere (places where it sadly wouldn't be that welcome, on the other hand, I'm afraid), but there's an even bigger argument laying down there within Schmerzkaufen's diatribe -- by offering an imperfect-though-palatable emulation for so many years despite having the tools to get a much better approximation (call them "hacks" if you will) the people are endorsing it as if it were the original behaviour, so it goes against the "telling the history as it was" argument MAME Dev claim once and again. Especially when most of the timesthe issues aren't documented when you run the emulation or even in the emulator's texts. The "flag" system of course doesn't suffice, but there's isn't one for the CP-S issue, for one. So it's fair enough that they advice against using hacks and offer a "pure" version of MAME so that proper emulation gets done some day (despite they admitting it won't happen for reasons) but persisting on feeding this unnatural dichotomy between devs and users isn't helping the project. MAME may well be the very first cause of the Retro Arch phenomenon.
schmerzkaufen:
Personally I'm fighting that dichotomy but I doubt anyone's getting my positioning. Specifically I disagree with the careless hacking camp on the one side and the too-final radical no-no-no mamedev's on the other.
In short I'm in favor of seeking compromises, like more warning messages, indeed flags wehre possible, but I'm thinking more of a an in-game OSD feature that displays how the gam's running and what's active along with it during gameplay.
Like we press F11 but we only have that poor refresh speed % meter, my idea is that it could show more information, like if lag reduction is used, cpu or refresh sliders, other 'cheats' like autofire or savestates, etc.
Imagine you press F11 and this appears, just to illustrate;
[SPEED100%, FD 8, CPU 74%, AF]
translated: game running at its normal refresh speed, frame_delay on set to 8, CPU underclocked @74%, AutoFire ON
Of course ideally the band would be moveable around the gameplay screen area maybe with smaller fonts too, in order to not get in the way.
This would also be a new and ideal way to confirm the validity of some videos, if for instance a wolf~ish variant of it would make that on-screen ingame info always-on
A build like that, unless it's stolen by RA again, would be a sort of permissive and restrictive one at the same time, useful for tweaking and cheating via hack~ish solutions, but also able to show that you're not using any - or specifically which one - during your run recording.
I'm convinced that permitting anything tweak/cheat/hack without care nor warning is damaging of course, goes without saying, but forbidding all this stuff absolutely and not pay attention to the fact that it makes people lose patience and interest then opt for the worse solutions is a huge mistake too, no doubt about it we're seeing the results today.
A tentative name for an hypotetical build: compromame? diplomame? middlemame? ;D
EDIT another illustration for fun, with another more complete info band;
[diplowolfmame 0.203 : SPD 101%, FD 7, CPU 60%, AF B1:3Hz, REC 1]
Calamity:
--- Quote from: Recapnation on November 02, 2018, 02:26:47 pm ---MAME may well be the very first cause of the Retro Arch phenomenon.
--- End quote ---
If MAME had better emulation of 16-bit consoles already, RA wouldn't be so relevant. I buy that part of the argument.
schmerzkaufen:
AND lag reduction, AND more complete saveable settings including those of the all-important shaders. People don't go to RA only because it's a mashup of many systems, but also a lot because it's got these features the official MAME don't, or only partially, or with longstanding issues.
I'm only interested in arcades and dislike several aspects of RA so for me the natural choice is GM, which you make reach significant milestones the one after the other in the right direction.
I would never consider the official MAME for playing, for me now it's almost like it's MAME that's just a necessary component of GM, and the latter is the real deal.
See what I'm getting at ?
Of course I know we owe a pharaonic amount of progress to mamedev's work since I'm following there too, but after I've had a taste of smooth scrollings, correct speeds, and low lag, it's impossible to play without those.
It's the same for many RA users, they've just made the wrong "emulator" choice where arcades are concerned, or they associate MAME to that unwelcoming rigid, spartan thing it tries so hard to be for sometimes good, but also sometimes really poor or not really honest reasons.
Also these users may have not heard of GM or they associate it to baseline MAME which they have already walked away from, which probably doesn't help.
Darn I'm at it again, I'm like Haze now, spamming the same story all over the nets. Anyway yes right, anything that improves the official MAME in the future in these particular areas, as well as consoles emuation of course, will make RA technically less relevant. Yet I can't help but think it's too late (crap I've already said that too, three or four times)
cools:
--- Quote from: schmerzkaufen on November 02, 2018, 03:19:34 pm ---but also able to show that you're not using any - or specifically which one - during your run recording.
I'm convinced that permitting anything tweak/cheat/hack without care nor warning is damaging of course,
--- End quote ---
Anything like this can be patched out and simply not displayed, if one really wanted to cheat in recordings.
--- Quote from: schmerzkaufen on November 02, 2018, 05:41:30 am ---
--- Quote from: Calamity on November 02, 2018, 04:28:25 am ---I believe I read that implementing wait states involves a non-trivial overhaul of MAME's architecture. It's not a matter of improving a couple of drivers. Based on that I guess that fixing those games without hacks means emulating the hardware at a much lower level, which also translates into more cpu intensive emulation. We can bet it'll be done at some point.
--- End quote ---
I suspected it was something heavy in any case yeah. An important overhaul probably won't happen just for better timings emulation anyway.
--- End quote ---
If someone were to put up a bounty for fixing CV1000 properly, I would personally stump up decent cash and I'm fairly sure there would be others that would do the same.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version