macmame converted inps for playback in pc-mame

[LN2]

Rick your urls are dead ends
apple something or other

There were at first.....for the 2 -fast ones cuz that made the filename too long for transfer I guess...but should be fine now.
Let me know if they still are failing.

My first sighting of one Ln2 recording
I request millpac 321,000 or whatever ,
first place .
My personal best is 70,000
on that rom.

I likely won't be able to convert that one. I tried with millipede some which should be similar. It's harder cuz it requires adding 8 bytes at the point the frame lengths apparently change from 40 bytes/frame to 48 bytes/frame.

I think even if I get that right, the pseudo-random choices in 0.60 would be different..so initial mushroom setup etc. would be different...thus causing a sync problem. millpac was done in 0.71.2.1 so it might be different. Odds are though the initial mushroom setup would be different.

[mahlemiut]

There are tons of PC inps recorded using versions of mame MOST don't have the old rom sets to be able to view. There are lots of xmame inps here too. MOST don't have linux installed to be able to view the xmame inps.

XMAME is mostly compatible with MAME. That's why we keep trying to get you to use it. It's also not a "pc-mame", as it can be compiled on various versions of Unix on various non-x86 systems.

If MacMAME Plus is based upon MacMAME, you may have some difficulty getting the inps compatible. At the very least, CPU cores must match. The amount of ASM optimisation in MacMAME may be another issue.

[Chad]

I realize in your and many others "dreams" you want me to just say screw macmame and go buy a cheap PC and trash all of my scores and play them all over again using pc-mame/wolfmame. That's not going to happen.

Not true, i was just not going to confirm a mac recording, i.e. if you submited it as a pc recording and had a "link" to the mac recording (instead of the other way around), i'd confirm it, it's just a question of being strict to the state of confirming recordings. You should only confirm recordings that you've played back, not modified ones. This is no way does screw mac mame recordings, it just still reiterates there are more pc users than mac users. I guess it's not much of a difference, just technically slightly different. I just don't feel right confirming an entry because of a link to it, i'd much rather confirm a recording that is ON marp's site not a link to it somewhere else.

[LN2]

XMAME is mostly compatible with MAME. That's why we keep trying to get you to use it.

Ok, but that's running on the same CPU.

if ported to Mac CPU, something like compiling it to use the CPU in little endian mode might be necessary. I have no clue how to do stuff along those lines. PPC-CPUs can run in 2 modes. There is big-endian mode which is how most software for Macs runs. There is also little-endian mode. Software like VirtualPC runs in this mode which really assists for emulating a PC.

It's also not a "pc-mame", as it can be compiled on various versions of Unix on various non-x86 systems.

ok, here's a question then. Are inps resulting from xmame compiled and running on a system like a Sun workstation compatible with xmame for x86? That's separate versus it just being able to be compiled.

I would certainly try xwolfmame built for OS X or even just regular xmame on OS X if it had a direct binary to run(versus a list of 50 instructions saying you must build this and this and that and get these libs and should have these libs etc. only to find out the instructions aren't entirely correct and builds result in several errors etc.

I don't want to have to learn how to program in linux to be able to build xmame. The Apple Dev tools for OS X might be able to do this stuff, but then I wouldn't be able to follow the instructions and would have to figure out those 50 steps.

A couple people had built xmame for x-windows for OS X with a older version...think it was 0.61...said it ran dreadfully slow where it wasn't useful. ie. 25 fps for pacman on a mac 3 times faster than what I have that gets about 180fps for pacman in macmame.

I'm not going to sacrifice a factor of 10 in speed.
I have no clue how much of that is from the assembly optimizations in macmame versus graphical hardware acceleration etc. The latest x11 implementation for OS X is much better than what I can use with my current OS.

I would prefer an xmame build that doesn't require x11. Maybe something like that could be built that doesn't require x-windows...and could even use the macmame graphical code to have the opengl acceleration etc.

It was this kind of thing I am tentatively planning on starting this summer. ie. take xwolfmame...and add to it what is necessary so it runs within OS X. The first version perhaps just CLI through the terminal...with later version with a GUI.

If MacMAME Plus is based upon MacMAME, you may have some difficulty getting the inps compatible.

Not sure why MacMAME+ is mentioned here. What's Macmame+ have to do with any of this?

All of the inps I recorded in macmame+ playback fine in the official macmame.

What I am referring to above is taking the xmame and macmame code...and first going through the code so the inp file format and contents are identical to xmame but using the macmame video and audio and controller sources at least..and get a build that runs and see how fast that is. If really slow then I would start to add assembly optimizations for certain game CPUs retesting the replay compatibillity to see if any changes from that etc.

Then start checking games and for those with differences like pseudo-random number generation finding a way to recode so that issue is gone. If within the assembly code, since I really don't know assembly I wouldn't be able to address those...except perhaps not use that assembly.

Does mame_rnd totally solve that situation?

Anyway, the result would be a different macmame build for OS X that users can download and directly use without having to install 5 other separate software packages and/or libs that yields replay files compatible with the pc version.

For me to even remotely build xmame or xwolfmame that would use hardware graphical acceleration I first would need to buy and install the latest OS, Panther(10.3.x) then run the updates on it to 10.3.2. I then would have to install the dev tools/x-code which has the gcc compiler and libs etc. as part of it.

I then would have to download Apple's latest x11 implementation ad install that.

I then would have to get and install fink. I then would have to get the SDL code and build the SDL libs...whatever they are. I then might have to get other libs also and install those.

After all of that, I then could attempt a make...although to install all of that stuff I first would need to get a larger HD.

[LN2]

I just don't feel right confirming an entry because of a link to it, i'd much rather confirm a recording that is ON marp's site not a link to it somewhere else.

If I submitted a pc converted inp for a game....then copied the inp link somewhere...then deleted that submission, it would revert to my mac listing right? The other inp would also still be here...but now on the MARP site right? ...so then I could put the MARP URL to the PC-converted inp in the description.

The net result of that is I get the PC one on the site, and preserve my mac submission and date etc.

Would this work?!? If that is how delete works then it should work....cuz when you delete it just removes the score from the scoreboard right? The inp archive isn't deleted.

Let me know. This can at least give you that level of comfort having the pc conversion also on the MARP site.

[tar]

Ln2 the urls in the original post work.
Picked up joyman.
I have mame32 version 67
windows ' 98
puckman rom
I looked at it
311,100 good work.
I would confirm it , but do not want to make skito mad .
Sorry for being cranky earlier (buttermaker)

[kranser]

Rick's Joyman-fast 1st place plays back in WIN MAME 67 to 618480 - not the 615,780 stated!

Kranser.

[LN2]

Rick's Joyman-fast 1st place plays back in WIN MAME 67 to 618480 - not the 615,780 stated!

Kranser.

yep, I mention that in the readme included.

I would have to watch it really closely to figure out where the heck the extra 2700 points comes from.

I said despite that I'll keep the score at what it plays to in macmame since that is where I played it....only fair.

[Chad]

yeah you're right rick, we don't have any sparc xmame inps to see if they work the same, there may be endian issues and more, but i'm surce they've compiled xmame on a motorola so xmacmame seems possible.

about my confirming naziism, you can't preserve the mac time submission, so here's a compromise, in your mac submissions (along with the link to the pc inp) provide the method by which you converted the mac recording to the pc recording, that way someone could actually do the same modification if you're link gets broken, and then i'll confirm it (or tar will =).

[LN2]

yeah you're right rick, we don't have any sparc xmame inps to see if they work the same, there may be endian issues and more, but i'm surce they've compiled xmame on a motorola so xmacmame seems possible.

Oh, I have no doubt it would compile for x-windows in os x. A couple had done it with 0.61 as I previously stated.

However, that means little if it doesn't help the replay cross-platform compatibility issues. Yes, it would auto-fix games like pacman where the difference is just from the deviation from the actual mame code in macmame for the inp format.

I could suggest this specific change for the pacman engine in the next macmame release so that at least would be fixed but the developer doesn't generally cater to wishes like that...and has in the past expressed that he doesn't care the file format isn't the same. He intentionally removed the header cuz he felt it wasn't necessary. He's not about to put it back.

That's a shame.

about my confirming naziism, you can't preserve the mac time submission,

so the "trick" I asked above wouldn't work? I thought when you delete a score the inp remains and if that player had a previously submitted score that one magically reappears on the scoreboard....right?

if so, then as stated above I can do a cat chow-chow-chow type thing where I submit the pc converted inp...so the inp gets named and stored at the MARP site...then delete that score...and my previous mac one would be back in the listing right? I can then add the URL for the "deleted" converted one to the description.

so here's a compromise, in your mac submissions (along with the link to the pc inp) provide the method by which you converted the mac recording to the pc recording, that way someone could actually do the same modification if you're link gets broken, and then i'll confirm it (or tar will =).

Well, if the above "trick" wouldn't work, then I would prefer to just add that conversion info to the description. It would only be a couple lines.

ie. Open up inp from a pc mame for the game. Copy first 112 bytes and paste at start of the macmame inp....save...done.

For even new scores, I likely would convert after the fact so wouldn't want to always have that nuisance of doing that quickly so the dates of my submissions are accurate.

[mahlemiut]

There are no file format differences between MAME and MacMAME other than the missing header. It is nothing more than a log of the games input port states.

Also, remember jmd? I'm pretty sure he had a non-x86 system, and I had no issues playing back his bubbles inps.

[LN2]

There are no file format differences between MAME and MacMAME other than the missing header. It is nothing more than a log of the games input port states.

The differences go beyond that. To convert the pacman clone replays, it's not just adding the header. You have to add 2 frames of data at the beginning to get the time in sync. Yes, the technical frame format is the same....but when it starts writing input to the file must differ. I still call that a difference in format. For joyman, where I used fs 6, I had to add 3 frames. I guess it's still just added 2 read frames of input.

For millipede and marble madness, to prevent those from going into service mode on playback, I have to add 8 bytes at the point where the frame size changes in the inp from 40 to 48 bytes per frame in 0.60.

Adding it anywhere else still results in it just going into service mode....so there must be some difference in mac vs pc mame code there for the inp format....the difference of 8 bytes...plus likely some timing buffer frames to add there also.

[Weehawk]

New confirmers are sometimes very enthusiastic and go on confirming sprees just for the sake of confirming.

Every recording at MARP that plays back should be confirmed. Every one that doesn't should be deleted. If that's what you mean by "just for the sake of confirming", then you are correct in my case.

Also, as a member/confirmer/Regulation Coordinator there's not that much I can do to encourage other members to submit using wolfmame for security and playbackability reasons. One thing I can do is to put forth an effort to make sure their submissions get confirmed. If someone can't be bothered to submit in wolfmame, I'm not going to be as concerned. You'll see that almost all of my "spree" was wolfmame and alphamame recordings.

when a recording doesn't work it would just waste the next confirmer's time if you don't post about it here.

Exactly.

Maybe with that inp it's a combination of a few things. Somebody who started watching it at normal speed just quit MAME after a while because he started getting drowsy. Sombody who watched it until the end (may God bless him) wasn't a confirmer.

I watched it through the sixth key only, and I wasn't a confirmer at the time.

I think those people are in the vast minority. This is the MAME Action Replay Page. It's not meant for people to simply list their scores. People come here to watch replays.

And how many people see it otherwise isn't the issue. It is the mission of MARP, as the owner purports (and he should know), to provide a repository of recordings for the purpose of allowing visitors to watch how a game is played by a master.

From the MARP Introduction Page:

Exhibitionism, in that I'm asking you to show off your high scores on the games, since it actually shows how the score was reached, instead of just being a big number.

From the MARP Rules Page:

MARP's goal is to emulate the experience of "watching a virtual master play an arcade game"

But perhaps Brian McLean summed it up best when he wrote:

MARP isn't for pussies too afraid to lose to show how they do anything.

If I could get a cool MARP logo for the front of a t-shirt, I'd put that quote on the back.

If people want to list their scores they should go to TG

Or simply put up their own page. "I have scored the following on the games listed:"

[LN2]

And how many people see it otherwise isn't the issue. It is the mission of MARP, as the owner purports (and he should know), to provide a repository of recordings for the purpose of allowing visitors to watch how a game is played by a master.

Oh, so those scores like 10,000 on Centipede are masterful? visitors will want to watch those ehh? IMHO they really shouldn't remain in the repository. However, MARP wants to be a scoreboard also...so keeps them all.

BTW, who said all visitors are using wolfmame or regular pc mame? What if mac visitors stop by? if they see some macmame replays it might encourage them to participate also. As stated above, most can't view most replays here anyway. The typical visitor would likely only be using one of the latest versions of regular mame or wolfmame. They wouldn't want to install 30+ versions of mame and have several rom sets for games that have changed in different versions of MAME which is what would be necessary to playback most inps.

MARP isn't for pussies too afraid to lose to show how they do anything.

I don't play in macmame to avoid having my inps watched. if I didn't want them watched then I wouldn't submit them here. I use macmame cuz a mac is all I have at home to play games on.

If I had a PC believe me I would be submitting using wolfmame versus putting up with all the crap I get from people just cuz I use macmame. They do get to see several of my efforts...but not all...no different than any long-term pc player that likely has submissions using versions like 0.35. No typical visitor is going to have 0.35 and the rom sets for 0.35...yet lots of masterful replays are using 0.35 or older versions.

For all I know some of the members here likely have access to a mac somewhere...perhaps school/college or the public library etc.

[LN2]

double post removed...