AlphaMame Blocking Bonus

Discussion about MARP's regulation play

Moderator: BBH

AlphaMame Blocking Bonus

Poll ended at Mon Mar 10, 2003 10:57 am

[For Blocking] <br>Block regular dos and windows mame recordings from beating alphamame recordings. All recordings prior to date of begining of the vote (3-3-2003) will be grandfathered to be "ok" to remain beating alphamame recordings.
24
80%
[Nay Blocking] <br> Leave it the way it is, all recordings can beat any others, even when regular mame can use cheating that is not detectable. But Allow for a tgmame leaderboard bonus to be decided later.
6
20%
 
Total votes: 30

LN2
MARPaholic
MARPaholic
Posts: 1669
Joined: Wed Jul 24, 2002 4:46 pm

Post by LN2 »

Yes, mame60 is considered a dos mame, use omame60 if you are recording on a mac.
Well, I have submitted a few new scores since March 3 using "mame60" as the version cuz the inp was compatible with mame60.

I see I can edit them to reselect omame60 so will have to do that for those few scores.

For games where the inp is compatible with pc-mames if I put that in the description is that enough for some pcmame confimer to confirm the score? I like it when scores that are possible to be confirmed are....would listing them as "omame60" make most confirmers overlook it and result in not getting confirmed? I guess that is my only choice.
performance is negligable if you're using a fast computer (one that actually will get 100% on games) so if you're using a slow computer/emulator then that's a problem right there. 90% speed acceptance for alphamame recordings is PRETTY leniant.
Yes, but as stated above for games where you already are using something like frameskip 6 in regular mame to reach that speed the higher CPU requirement for the real-time encryption as you play would slow this down and perhaps force you to go to even a higher frameskip.

As you see there are even some games where even if you have a somewhat fast CPU you can't play at speed without using frameskip or turning off the audio.
Encryption has to be done during recording for more security, i.e. you could stop mame before you escape and copy your cheated inp right before the alphamame encrypts your recording, besides i don't think will not buy ENOUGH cpu cycles to make a difference.
Well, you lost me here. How do you exactly "stop mame" and copy the inp file elsewhere? Do you mean when you actual game was over you just shell-out to Windows while alphamame is still running and replace that inp file with a different one..so when you hit escape to quit alphamame it encrypts that "cheat" inp instead? How would that be possible? The original inp file would be in an "open" status with alphamame still running. You wouldn't be able to rename or replace it with anything cuz the file is busy/in use. Don't tell me Windows allows you to replace an open file?!?!? I have used PCs and Windows a lot over the years but never had tried it to even know if possible. That certainly isn't possible on a mac. It's something to test with a regular mame to see if windows allows that. I thought you get a "file is in use" error when you try to do replace or move or delete a file in use.

If alphamame has that inp file in "open" status the entire time, then when you hit escape quickly closes it then reopens that same path and filename for encryption, you wouldn't be able to do what you state above.
Yes, you would be able to copy that alphamame inp elsewhere at that point to have an unencrypted form of the inp also before you hit escape and it gets encrypted and the original deleted...but so what?

Does Windows let you move a file when busy/in use? If so then that wouldn't be good cuz then in that close/reopen it could have been open to a file you moved but then in reopening that sets the path back to the inp folder where you might have a different one there now....cuz you moved the original....again would need to test if Windows would allow that...I would hope not. You shouldn't be able to do anything with a file that's busy/in use except copy it.

One thing I see as not the best here is even for games that are cross-platform compatible other platforms can't watch those encrypted inps. The long-time request for cross-platform compatible inps is moot cuz the encryption.

This is why a month ago in one of the alphamame threads I suggested why doesn't alphamame produce the normal inp file but then a .enc file that is generated by alphamame also that would be required to confirm the inp file back in alphamame. If the inp file was altered in any way then it wouldn't "jive" with the info in the .enc file so alphamame would report this etc.

That way, we still have the basic inp file to share with others to view with different variants of mame of that same core version. Part of MARP's misison statement I thought was to share recordings of each others gameplay. Part of that is lost now with alphamame for those not running M$Windows.

Many liked that idea when I gave it a month or so ago...but seems to have been forgotten. That also removes the filesize problem cuz as Barry stated the alphamame encrypted inps don't compress nearly as well as regular inp files do. That's when I suggested just have the raw inp file but a .enc file etc. that would confirm the inp has been unaltered plus have the other stats like average speed/framerate, frameskip used, and whatever else alphamame provides.
The emulation for some games is going to be at least a 100 times more needy for processing time than the encryption.
Ok, so are you saying it only makes a difference of 1% in the time to do a frame? I had asked above how much of a performance hit are we talking here. If only 1-2% type of thing fine...no biggie at all, but then I wouldn't have expected to see posts by other players saying how much slower a game runs in alphamame versus regular dmame/mame/mame32. The fact they are posting about it the performance hit must be fairly significant.

I don't think my above suggestions comprise the security of alphamame at all.
User avatar
mahlemiut
Editor
Posts: 4186
Joined: Mon Feb 04, 2002 10:05 pm
Location: New Zealand
Contact:

Post by mahlemiut »

LN2 wrote: I request again that "mac60", "mac58", and "mac53" be added to the version select menu on the submit page. That way Zwaxy can have the code not treat those inferior to alphamame inps. ...guess he will need to do same with a few xmame builds also.
I've added mac60 to the versions list, along with alphamame 0.66. I tried to add mac58 too, but something kinda went wrong, so I removed it and MARP worked again. :oops:

EDIT: Sorry, mac58 is there too, I just guess caching got in the way of my fuck-ups. :)
- Barry Rodewald
MARP Assistant Web Maintainer
Image
User avatar
Chad
Tournament Coordinator
Posts: 4463
Joined: Tue Mar 05, 2002 3:15 pm
Location: calif

Post by Chad »

Thanks barry for the mac versions, hopefully people will be able to wait for the mameversion list to be downloaded :)

Ahh, there's a Rick post i'm used to seeing! :)

Ok: as far as speed goes alpha mame theoretically it would be nice to playgames at 100%. And the regular marp site is giving you a 10% advantage, (tournaments give you a 5% advantage but they usually choose fast games). So if there is a performance degrading part about alphamame it would have to make it 10% slower than regularmame and that just CANT be the case for the average game, i could see pacman or a simple game actually using only 9 times the cpu time that the encryption does (this can be tested by playing back inps in no throttle mode for mame and alphamame to see what the fpses are in both playbacks). If pacman playsback at 450% with alphamame and 500% with regular mame then alphamame has a 10% degrading factor, but i'd imagine those %'s to be pretty close.

And for the encryption, yah YOU can't do the copying of files while they are being created but hackers CAN. you have to be really paranoid when it comes to security or nothing will get done :) and again I really think it's not important to gain a 10% (at most) difference. Actually Try alphamame and then regular mame with your emulator and see if they playback recordings nothrottle at a similar fps (with the same game and same frameskipping), that would be the biggest test, if they are actually way different speeds barry and i will look into it but i don't think they will be significantly different.
-skito
LN2
MARPaholic
MARPaholic
Posts: 1669
Joined: Wed Jul 24, 2002 4:46 pm

Post by LN2 »

mahlemiut wrote:I've added mac60 to the versions list, along with alphamame 0.66. I tried to add mac58 too, but something kinda went wrong, so I removed it and MARP worked again. :oops:

EDIT: Sorry, mac58 is there too, I just guess caching got in the way of my fuck-ups. :)
Yeah, browser caches can often be a nuisance....although if like mac web browsers you can alt-refresh to force a total refresh ignoring the cache so you can see the updates.


Thx for adding those versions Barry. I think I'll go through all my submissions that have other 0.60 selected and change those over to macmame60.
LN2
MARPaholic
MARPaholic
Posts: 1669
Joined: Wed Jul 24, 2002 4:46 pm

Post by LN2 »

Chad wrote:Ahh, there's a Rick post i'm used to seeing! :)
..anytime. :D
If pacman playsback at 450% with alphamame and 500% with regular mame then alphamame has a 10% degrading factor, but i'd imagine those %'s to be pretty close.
Well, if you think that's really about all the hit you would take then fine.

Pacman isn't a good test though. Given my above you need to test this on a game in regular mame you can't even get 60fps for with throttle off and frameskip set to 0....then see how much slower it runs with alphamame. You would think it would be that same 10% kind of thing....but would be a good check.

I can try this in VirtualPC on my mac I guess...although even pacman in VPC struggles to run at full frame rates...shows you just how slow the emulation is...although my mac itself nowadays isn't exactly fast itself. :P Anyway it could be a good judge of how much of a percentage for each game frame/tic the encryption stuff in alphamame adds to it.

Actually, can't you tell within alphamame directly? If you are playing using alphamame and not recording, then you aren't using the cpu for encryption...or is alphamame where you are always recording for the game you try and play? I guess I'll have to try it in VPC.

If I had other reasons to need a PC I would get a fairly cheap one and use that for my mame play...but that would really be about the only reason to get a pc for me right now...not worth even $500 just for that.
User avatar
Chad
Tournament Coordinator
Posts: 4463
Joined: Tue Mar 05, 2002 3:15 pm
Location: calif

Post by Chad »

LN2 wrote:Pacman isn't a good test though. Given my above you need to test this on a game in regular mame you can't even get 60fps for with throttle off and frameskip set to 0....then see how much slower it runs with alphamame. You would think it would be that same 10% kind of thing....but would be a good check.
Remeber as you choose a slower game than pacman the alphamame slowdown factor is going to get BETTER. the slower the game gets the better ratio from emulation to encryption will be. so pacman will be the max slowdown alphamame will give you compared to regular mame.

also, alpha mame only encrypts when recording so it should act like regular mame when playing regularly.
-skito
LN2
MARPaholic
MARPaholic
Posts: 1669
Joined: Wed Jul 24, 2002 4:46 pm

Post by LN2 »

Chad wrote:Remeber as you choose a slower game than pacman the alphamame slowdown factor is going to get BETTER. the slower the game gets the better ratio from emulation to encryption will be. so pacman will be the max slowdown alphamame will give you compared to regular mame.
Ok, for games with really high CPU requirements I understand what you are saying...but it's only those games where the CPU overhead for the encryption really matters. Using your hypothetical example of pacman running at 450% vs 500%...doesn't matter cuz you play it in both at 100% speed with frameskip 0 and get identical performance.

For games with high CPU requirements yes the amount of CPU used for encryption for each frame is roughly same as it was for pacman...so actual %overhead for that has dropped...but now that amount it has dropped reduces actual gameplay performance..where in pacman it doesn't reduce gameplay performance at all.

...so how much of a hit is it? Your 450% vs 500% for pacman seemed hypothetical. Was that a real test on your PC? I was just curious how much is it cuz a few had already posted about performance issues...so that would tell me it must be fairly large....perhaps more than your 10% hypothetical example.

I guess in most cases for most games it doesn't matter and for even some where you need frameskip 4 or 6 or 8 you can always notch that up 1 more frame skipped to compensate....so perhaps adding that all together it isn't worth changing?

I was hoping it more for the aspect of having the original inp file still saved...with the separate .enc file...cuz now most new uploads will be with alphamame meaning my only chance of watching even those with inps normally compatible with macmame would be watching them in VirtualPC. It's generally painful to try and watch even my own inps to see if they are cross-platform compatible. I normally get it started then go do other stuff and come back in an hour or so to see the final score is ok. I test most of them with no audio and frameskip 6 to even get about 50% game speed for the playback for a game like Cameltry. Even if the macmame code was changed so inps are cross-platform compatible for all games sometime in the future it wouldn't mean anything cuz you still wouldn't be able to view any current alphamame encrypted inp files.
User avatar
Chad
Tournament Coordinator
Posts: 4463
Joined: Tue Mar 05, 2002 3:15 pm
Location: calif

Post by Chad »

did do a real test, only hypothetical numbers, a real test for me would be in the 2000% range for pac man :)

Themoretically it's possible to decode an alphamame inp file to a regular mame inp file, with out a garuantee it would actulaly play back in the regular mame version, let alone a mac mame version. However being paranoid, this decryption program could be a tool that hackers could use to backwards engineer the encryption.
-skito
User avatar
Barthax
MARP Seer
MARP Seer
Posts: 691
Joined: Fri Sep 27, 2002 1:13 pm
Contact:

Post by Barthax »

LN2 wrote:
Encryption has to be done during recording for more security, i.e. you could stop mame before you escape and copy your cheated inp right before the alphamame encrypts your recording, besides i don't think will not buy ENOUGH cpu cycles to make a difference.
Well, you lost me here. How do you exactly "stop mame" and copy the inp file elsewhere? Do you mean when you actual game was over you just shell-out to Windows while alphamame is still running and replace that inp file with a different one..so when you hit escape to quit alphamame it encrypts that "cheat" inp instead? How would that be possible? The original inp file would be in an "open" status with alphamame still running. You wouldn't be able to rename or replace it with anything cuz the file is busy/in use. Don't tell me Windows allows you to replace an open file?!?!? I have used PCs and Windows a lot over the years but never had tried it to even know if possible. That certainly isn't possible on a mac. It's something to test with a regular mame to see if windows allows that. I thought you get a "file is in use" error when you try to do replace or move or delete a file in use.

If alphamame has that inp file in "open" status the entire time, then when you hit escape quickly closes it then reopens that same path and filename for encryption, you wouldn't be able to do what you state above.
Yes, you would be able to copy that alphamame inp elsewhere at that point to have an unencrypted form of the inp also before you hit escape and it gets encrypted and the original deleted...but so what?

Does Windows let you move a file when busy/in use? If so then that wouldn't be good cuz then in that close/reopen it could have been open to a file you moved but then in reopening that sets the path back to the inp folder where you might have a different one there now....cuz you moved the original....again would need to test if Windows would allow that...I would hope not.
Not sure about 95, 98 & ME, but under NT, 2000 & XP you can rename a file that is currently open & in use. Also, with a full-screen CLI / DOS mame, ALT+ENTER will stop the emulation & return the user to Windows. MAME will appear in the task bar & can be clicked on to continue from where it was left off (DOS MAME may often crash in this scenario, however).
LN2
MARPaholic
MARPaholic
Posts: 1669
Joined: Wed Jul 24, 2002 4:46 pm

Post by LN2 »

Barthax, yes but that file even if you rename it will still have the same pointer within the running alphamame pointing to that file. Once opened files are all just pointer locations..so the actual filename no longer matters...which is why you can change that name even for files in an open status. BTW you can do that on macs also.

You make a good point though that if you formally closed the file then reopened a certain filename you could have renamed that original inp..then put a different one there..then return to alphamame. That would bypass the security if alphamame formally closed then reopened that file cuz then the filename does matter for opening.

You can easily avoid that if necessary using a sequential file format instead and rewinding to the start of it to do the final encryption process. Then the pointer is always to that same inp file regardless what you have named/renamed it.
User avatar
mahlemiut
Editor
Posts: 4186
Joined: Mon Feb 04, 2002 10:05 pm
Location: New Zealand
Contact:

Post by mahlemiut »

You can always bypass that by killing the process before it gets a chance to start encrypting.

Also, under Win32 and Unix/Linux, the window losing focus or being minimised by default shouldn't pause the game, at least for long. It's also possible to do this with DOS MAME too.
- Barry Rodewald
MARP Assistant Web Maintainer
Image
LN2
MARPaholic
MARPaholic
Posts: 1669
Joined: Wed Jul 24, 2002 4:46 pm

Post by LN2 »

mahlemiut wrote:You can always bypass that by killing the process before it gets a chance to start encrypting.
You lost me here...cuz if you did that...then on playing it back in alphamame it would alert you it doesn't match and wasn't encrypted wouldn't it?
User avatar
mahlemiut
Editor
Posts: 4186
Joined: Mon Feb 04, 2002 10:05 pm
Location: New Zealand
Contact:

Post by mahlemiut »

Not as it is, it doesn't, that's why it encrypts on the fly.
- Barry Rodewald
MARP Assistant Web Maintainer
Image
kranser
MARP Knight
MARP Knight
Posts: 333
Joined: Wed Mar 13, 2002 11:23 am
Location: London, UK

Post by kranser »

Does that mean you can play back non-alphamame recordings in alphamame? (i.e. win66 recordings play back in alphamame66?) if so, great! Thanks.

Kranser.
User avatar
Chad
Tournament Coordinator
Posts: 4463
Joined: Tue Mar 05, 2002 3:15 pm
Location: calif

Post by Chad »

no, because it decrypts on the fly too, so it'll try to decrypt an unencrypted inp file, and results in lots of credits and buttons being pressed when you didn't realy do so.
-skito
Post Reply