forums.marpirc.net

I remember way back (.72 cycle) Luja mentioned that this game had gotten a settings change--that would be from Easy to Normal and bonus lives from 10/30 to 20/40--maybe other things.

I also noticed in MAMEInfo that this game's MCU has now been properly decapped, but the changes are not yet in effect.

Finally, you'll notice a lot of sarcastic comments in the descriptions for MARP's .inps about this game being 'a cure for insomnia', 'i'm so bored', etc. etc. You'll also notice that VERY FEW .inps were made on this game post .72.


IMO, these three factors make this game an excellent candidate for a score wipe and rules/settings lock with the new settings when the decapped MCU is officially applied, which i'm guessing will probably be in .140. opinions?

Hello.

Can I ask what is an MCU and how his emulation is going to change gameplay of the game?

Thanks

If I remember my comp. sci., it's basically a bridging microprocessor.

I think the key answer would be how this would affect gameplay, I don't know and I'm assuming QT doesn't know since he's too bored to analyze the recordings :) Cronos: I think you could answer this yourself by noticing a difference in play from your most recent to your oldest recording. Or maybe there won't be changes until the decapped MCP gets applied in a future mame version. If or When that happens, then you should revisit a score revamp.

At this point, the MCU isn't hooked up in MAME, it isn't known at this stage just what it controls.

MCUs will typically take care of a specific job, anything from handling input, coins, or providing some form of protection.

actually, my understanding is a bit fuzzy, because i thought MCUs were used to communicate between different processes--like a manager or moderator of sorts. i thought, and i could be wrong here, that they were often used to synthesize a 'protection' element with regular gameplay--but Barry knows better than i do. as Barry said, the changes are not yet in effect--but the main point, skito, is that the dip settings HAVE changed and easy was a ridiculously low setting to start with (this is a game i have actually played in arcade)

MCUs are basically simplified CPUs, they usually use a small amount of code and RAM.

Old PC keyboard controllers are MCUs (another common use for them).

My old recordings are from ten years ago! , so first I have checked the actual default settings of the game and his clones.

All are set to 3 lives and normal difficulty, the same than delfault machine and twin galaxies settings. http://forums.twingalaxies.com/viewtopi ... 96&t=18542

Then I have downloaded an old mame version (alphamame32 0.70) and checked the defaults again. Q.T. Quazar is right here because all were set to easy difficulty.

But the worst thing I have seen is that with this version of mame, the game is something like four times slower that the actual game speed. With the f11 key I have checked that game goes at a constant 60fps, but gameplay speed is ridiculously slow, I think that with this speed I can play for a week without loosing a live.

I think the key is here: 0.112u4: Aaron Giles removed MDRV_CPU_PERIODIC_INT in the Street Fight machine driver - no more slowdowns.

So I have tried another mame versions with the following results:

wolfmame 0.106 - Game slow
wolfmame 0.111 - Game slow
wolfmame 0.112 - Game broken
wolfmame 0.113 - Game at normal speed

The new inps I have submited this days are recorded with wolfmame 0.138 and the game speed there is normal too.

Seeing that, I think that as a minimun, all pre 0.113 version should be banned.

I imagine the MCU handles the coins, there is no sign of any real 'protection' in the driver except for it being very fussy about the coin latch. IOW it won't make any real difference to gameplay, unlike in some cases.

Haze wrote:I imagine the MCU handles the coins, there is no sign of any real 'protection' in the driver except for it being very fussy about the coin latch. IOW it won't make any real difference to gameplay, unlike in some cases.

ah. quite clearly my understanding of the MCU's function was not quite clear of an understanding

Cronos wrote:My old recordings are from ten years ago! , so first I have checked the actual default settings of the game and his clones.

All are set to 3 lives and normal difficulty, the same than delfault machine and twin galaxies settings. http://forums.twingalaxies.com/viewtopi ... 96&t=18542

Then I have downloaded an old mame version (alphamame32 0.70) and checked the defaults again. Q.T. Quazar is right here because all were set to easy difficulty.

But the worst thing I have seen is that with this version of mame, the game is something like four times slower that the actual game speed. With the f11 key I have checked that game goes at a constant 60fps, but gameplay speed is ridiculously slow, I think that with this speed I can play for a week without loosing a live.

I think the key is here: 0.112u4: Aaron Giles removed MDRV_CPU_PERIODIC_INT in the Street Fight machine driver - no more slowdowns.

So I have tried another mame versions with the following results:

wolfmame 0.106 - Game slow
wolfmame 0.111 - Game slow
wolfmame 0.112 - Game broken
wolfmame 0.113 - Game at normal speed

The new inps I have submited this days are recorded with wolfmame 0.138 and the game speed there is normal too.

Seeing that, I think that as a minimun, all pre 0.113 version should be banned.

yes, i've had the same thought (thanks for the further check, btw). as i've been doing the win37 audit, a number of games have come up (particularly shmups) that cannot possibly have been in the 90% range for players 10 years ago... as my OCed Core2 is still choking on them (although, caveat: i haven't been doing fskip specific or nosound settings). i think the settings changes that happened mid-cycle on a lot of these games gives us a gracious way to do some housecleaning. but that's just my opinion.

i also think the 90% min speed rule should be bumped up now. i think i was the one that originally proposed it in '99 (since slowdown was a big problem then), but it seems antiquated now ever since MAME began to support multi-core. perhaps 97%?--that's enough to include 'accidental' variance, i think.

I have upped one recording to show things:

http://replay.marpirc.net/r/stfighta

I have reached 1 million points with the first live using a pre 0.113 mame version (wolf106)

If you take 1 minute to check this recording:

http://replay.marpirc.net/inp/7/f/9/jog ... olf106.zip

against this one, recorded with an actual mame version:

http://replay.marpirc.net/inp/f/8/1/jog ... olf138.zip

You will notice a HUGE difference in game speed between them. So in my opinion in all pre 0.113 version, the game is bad emulated, and must be a rule that ban the use of old versions of mame, with this game and all his clones.

I gonna look through this as quickly as I can.
I don't know the game yet, but at first sight you're perfectly right, and all premame113 inputs should be zeroed.
I need to test it myself first ...

Ok I've looked better : Cronos is right, speed is much slower before m113.
So yes, I think we can now zero all pre-mame113 inputs.
Is anyone against this ?
or can we proceed ?

It's tough to wipe 10+ scores. Perhaps for all scores that are wiped, we keep a link to their download able files on this thread. Then I think it would be good to wipe all slow playing scores.

wouldn't just zeroing them and leaving the link up be easier? players who are still active are going to replace them anyway.