"Preload" uploads

[mahlemiut]

Both term2la1 and term2la2 recordings by preload are now gone.

[preload]

The problem for term2la1 and term2la2 is the sound.
I've play this game with the sound activated and the speed has gone down. And after to submit the record i've not see the inp.
Excuse me for my error.
I've just upload a new record whit FS0 and No sound

[LN2]

a tip..if you aren't getting a steady 95-100% speed at fs0, then use some frameskip...like fs4 or fs6.

In most cases I would rather play a game with sound and fs6 at 100% than play it silent at fs2-4 or even fs0.

[SJK]

preload's point blank inp has quite a lot of slowdown at times... mostly staying at 80%....

[Weehawk]

preload's point blank inp has quite a lot of slowdown at times... mostly staying at 80%....

Does the game run slower at these points normally, or would it likely have been done deliberately?

[Buttermaker]

Does the game run slower at these points normally, or would it likely have been done deliberately?

If the game itself and not MAME runs slower the recording speed doesn't go down.

[Weehawk]

Does the game run slower at these points normally, or would it likely have been done deliberately?

If the game itself and not MAME runs slower the recording speed doesn't go down.

Guess I wasn't clear.

Does MAME normally run slower at these points in this game, or is it likely that some deliberate slowdown tool was used?

[Buttermaker]

There are games that slow down when the arcade hardware isn't fast enough. When that happens WolfMAME does not report a slowdown because it's part of the emulation.

In this case WolfMAME reported 80% speed. That means preload's computer wasn't fast enough for the game.

[Weehawk]

There are games that slow down when the arcade hardware isn't fast enough. When that happens WolfMAME does not report a slowdown because it's part of the emulation.

And this has nothing to do with what we are talking about.

In this case WolfMAME reported 80% speed. That means preload's computer wasn't fast enough for the game.

I understand that. But SJK reported that this was only on some significant places in the game. What I am asking is if it is reasonable to believe from the places in which it occurs, that this happened naturally as a result of the processing necessary, or does it seem likely that it was orchestrated deliberately to make those parts of the game easier?

And if it was unintentional, what then? The rules say:

g) Games must be played at a MINIMUM of 90% of their full speed. Use a framerate counter if you are unsure.

Does this mean 90% at every point in the game?

[Buttermaker]

And this has nothing to do with what we are talking about.

I said it because... ah, let's forget it or this will go on forever.

I understand that. But SJK reported that this was only on some significant places in the game. What I am asking is if it is reasonable to believe from the places in which it occurs, that this happened naturally as a result of the processing necessary, or does it seem likely that it was orchestrated deliberately to make those parts of the game easier?

I watched the replay and it happens only in scenes where the screen is scrolling. Preload's CPU is too slow.

And if it was unintentional, what then? The rules say:

g) Games must be played at a MINIMUM of 90% of their full speed. Use a framerate counter if you are unsure.

Does this mean 90% at every point in the game?

It means average recording speed as far as I know which is absolutely pointless. If a player gains an advantage because of a slowdown the average speed should not matter. The recording needs to be zeroed.

[Weehawk]

let's forget it or this will go on forever.

Forget what?

it happens only in scenes where the screen is scrolling. Preload's CPU is too slow.

So it wasn't done deliberately. Okay. Shit happens.

It means average recording speed as far as I know which is absolutely pointless. If a player gains an advantage because of a slowdown the average speed should not matter. The recording needs to be zeroed.

I take your point, but I'd feel a lot better about it if the rules were clearer.

[LN2]

I understand that. But SJK reported that this was only on some significant places in the game. What I am asking is if it is reasonable to believe from the places in which it occurs, that this happened naturally as a result of the processing necessary, or does it seem likely that it was orchestrated deliberately to make those parts of the game easier?

This seems obvious enough Weehawk. The CPU demands of the game vary in different areas of the gameplay. For some areas preload's CPU might be able to handle it so he gets 100% speed or very close to it...while in other areas it's not fast enough so the speed drops off to 80% or lower etc.

If that is the case, then he should find and use a frameskip value that allows the slowest part of the gameplay to still be at 90+%. I can't tolerate any variance in speed like that anyway..cuz it throws off timing etc. so I use frameskip so the stats show 100%...might drop to 95-100% here and there very briefly at the most.

Think of it like any other PC game you might play...like UT2003 for example. You might have areas you get 50+fps yet other areas you get only 20fps. This is because the processing demands of the game vary versus the scene you are in and how many polygons etc. and models and lighting entities etc. there are to handle.

This can easily unintentionally happen if someone just uses the stats and sets frameskip only with the initial demo/attract mode stuff running...then doesn't check it in actual gameplay. The attract mode for most games is often silent or nearly silent compared to actual gameplay...so even if the attract mode shows some serious action, it's without sound so would play at a faster speed than real gameplay with the sound.

So preload can choose to turn off the sound completely which would help a lot, or just use frameskip like fs6(which I like cuz skipping exactly half the frames still makes the action appear totally smooth. At odd frameskip values you get some jittery type motion in the vision likely cuz how it syncs.) with sound to play the game at 100%.

Depending on the game, he might require even more. I have a few games I have played where I have used fs8 or even fs10! hehe

I always playback my inps to make sure they playback correctly and also to check the playback speed. I think anyone not sure if they are above the 90% can playback their own inp and see if that is the case before submitting it.

And if it was unintentional, what then?

whether it's intentional or not isn't the issue for the direct consequences. You could actually make it slow by having stuff running in the background so less CPU is available to MAME..like playing mp3s.

Whether intentional or not the score would be zeroed out for slowplay.

Does this mean 90% at every point in the game?

this is a good question. I think very minor drops below 90% are tolerated as long as the vast majority of the gameplay is above 90%.

I'm not sure what wolfmame is reporting exactly when it shows the recorded speed. That must be some kind of average..but not sure over how many frames that is etc. without looking at the code.

[mahlemiut]

It reports exactly what MAME says the speed is. WolfMAME just reads it from a performance structure that MAME provides and writes the speed to each frame of the inp. The speed does average out to a point, especially when there is some slowdown, MAME will try to catch up.

PS: Can you provide a link to the MacMAME Plus site?

[LN2]

It reports exactly what MAME says the speed is. WolfMAME just reads it from a performance structure that MAME provides and writes the speed to each frame of the inp. The speed does average out to a point, especially when there is some slowdown, MAME will try to catch up.

ahh, so are you saying it's using the actual speed number reported in the stats part of MAME? ...then saves that to the replay file using a couple bytes?

ok, so the question still sort of stands then...hehe
What average is that speed number MAME shows? It can't be frame by frame....although if you turn throttle off you can see the number fluctuates several times a second.

...perhaps it's tied with frameskip which is based on a 12 frame cycle. Could the reported fps and speed be a 12 frame average?

PS: Can you provide a link to the MacMAME Plus site?

I got it from http://www.emulation.net/mame/

It has the macmame-GUI for it but with the plus features added...plus the code for 0.76.

I have only tried a few games with it on Sunday but it seemed pretty good.

Given the assembly code speedups aren't included though worries me a tad for performance for some game CPUs. Most classics seem to play just fine though. I actually get slightly faster framerates for pacman type games with this mame versus the official one. This might be because it was compiled with project builder and gcc instead of the official build which is done with Codewarrior. Plus I think the asssembly optimizations were removed from the official build for the pacman type games a few versions ago.

I will try and see if I can get this project to compile. If so, then I would only need the changed source code files from you to add the "wolf" stuff to it.

Odds are I would be able to fairly quickly apply those changes if there aren't very many. I noticed though with this macmameplus the inp file format still seems to be the same as the official macmame...lacking the header info for the inp etc.

In a nutshell, what does "PLUS" add to MAME? All I have noticed really is the "blue" colored config menus and stats you can view...and when you playback an inp it shows you when you reach the end of playback(which is cool).

Also, this version in the GUI has an option to only use the MC68000 C core to play. I have not tried that option yet. I figure it might run slower with it...but perhaps would resolve some inp compatibility issues with PC mames.

[Weehawk]

This seems obvious enough Weehawk. The CPU demands of the game vary in different areas of the gameplay. For some areas preload's CPU might be able to handle it so he gets 100% speed or very close to it...while in other areas it's not fast enough so the speed drops off to 80% or lower etc.

Yes, it is obvious. Unfortunately it does not address my question at all.

whether it's intentional or not isn't the issue for the direct consequences

It certainly should be.

You could actually make it slow by having stuff running in the background so less CPU is available to MAME..like playing mp3s.

Or simply run a program specifically designed for this purpose.
That was the point of my question.

Whether intentional or not the score would be zeroed out for slowplay.

If it were intentional, most of us feel that this would not be enough.

viewtopic.php?t=10971