Klax, Atari ST to Jaguar conversion

[sh3-rg]

Holy crap. I didn't mean to start all this

 

You better think more carefully about asking for games, young man. Thiiiiink! There's coders in them there hills!

[MAYAman]

I still find it hilarious when people get all up in arms about someone asking a simple question. Stop being so damned anti social. Not everyone is a coder or cares about what you think. If someone asks, "Hey would it be possible to have game X on the Jag" why get all up in arms? The Jag has been long dead and thanks to people like us that still play the games, its still talked about. Its not only coders that play it so get off your high horse if you're on one and realize that everyone here usually loves the Jag, and if they are a troll then pity them as trolling the forum of a long dead system is pretty sad. Enjoy your jag, be nice, and be civil.

[+Gemintronic]

Enjoy your jag, be nice, and be civil.

 

Is that like the piece of advice where you can pick two: fast, cheap and well done?

[LinkoVitch]

Sorry if my posts here upset people, I let this conversation get under my skin a little too much. My bad.

[sh3-rg]

I still find it hilarious when people get all up in arms about someone asking a simple question. Stop being so damned anti social. Not everyone is a coder or cares about what you think. If someone asks, "Hey would it be possible to have game X on the Jag" why get all up in arms? The Jag has been long dead and thanks to people like us that still play the games, its still talked about. Its not only coders that play it so get off your high horse if you're on one and realize that everyone here usually loves the Jag, and if they are a troll then pity them as trolling the forum of a long dead system is pretty sad. Enjoy your jag, be nice, and be civil.

 

Sarcasm. How does it work?

 

smh...

 

That was obviously a joke to JagChris and certainly none of your concern, but I appreciate you taking the time to grace us with your presence and judgement, if this thread was lacking in anything, it was a little dose of drama to brighten it up.

 

 

Yes, that too was sarcasm. Learn to spot the early warning signs and stay alert.

 

 

At first I guessed this was a wrong thread post because it seemed at odds with the topics being discussed here, as it's clear nobody is at all upset or annoyed that the OP created a topic about a great Atari game "Klax" for the Jaguar, specifically the ST version being refactored (as far as these kinds of threads go, this was a pretty damn realistic and a positive suggestion, and I'm kind of up on such topics for obvious reasons). But thinking back to a similar outburst, it's more likely your emotional outpouring is related to a a thread from years ago. Interesting you held on to the hurt so long. Maybe you could go dig up that old thread and begin posting in it again and work through your issues? Actually, reading it again, it's quite understandable why you didn't get on too well at GAF if you fail that badly at spotting a simple, friendly sarcastic comment from one regular forum poster to another, the kind of people who have been known to even chat privately regarding various topics relating to the Jaguar. My guess is you never made it past Junior Member status before rage quitting there or having a meltdown and getting yourself banned, right? The feels!

 

FYI, the turn this discussion took was in no way an indictment on the OP's question or its validity at all, what happened was it moved on to the subject of code profiling metrics and how they're as meaningless under emulation as they are away from real-world game conditions. If you want to offer your thoughts on these topics, go for it. If you'd prefer to hissy fit direct attacks on users maybe it'd be better to just move along to the next thread and refrain from that. Also, those condescending and trite non-moderator moderating pearls of wisdom you offer, the "can't we all just get along" kind of utterance, that's especially patronising... like you're somehow at the heart of what goes on around here and are suffering due to it when in reality you pop in here rarely and don't know one user from another.

 

If my posts are really that difficult to fathom and cut you so deeply and you have such a lack of ability to spot sarcasm when it rears its ugly head, I suggest for your own sanity and blood pressure levels that you put me on ignore. Ignorance is bliss, after all. Just pop my username into this box and hit Enter, it's your key to a happy and fruitful AA experience: http://atariage.com/forums/index.php?app=core&module=usercp&tab=core&area=ignoredusers

 

Anyway, thanks for dropping by, see you in another 12 months! Cheerio!

 

 

ps Thanks for all the helpful advice, if anyone here has an ounce of common sense, they'll promote your status to moderator by the end of the day.

 

[MAYAman]

sh3, with all due respect my comments don't just stem from that post, it is from watching the way others are treated on here by certain individuals. Also, last time I checked I'm not a swami and can't read peoples minds, let alone posts. So yeah, there are people here with rude attitudes and lofty opinions of themselves. Not everyone is a developer. Regular fans of video games like myself have questions too, sometimes silly.

 

It hardly required a long drawn out sarcasm laden response like the one you posted. Did it cut you or something? Sheesh.

Edited February 25, 2014 by MAYAman

[+Gemintronic]

I don't think anyone's being a royal nob. Different people interact, er, differently. I tend to soak it all in. Other people challenge and debate as a response to feedback. I think to an outsider it could look like an argument at times. I don't see any rudeness here myself.

[MAYAman]

OK, if it isn't an issue then just ignore.

[+Sauron]

This thread is getting close to being locked. Stop the bickering, please.

[VladR]

It's not a matter of a little variance in accuracy, it's orders of magnitude off-the-chart plain wrong and willfully misleading.

So now I'm somehow Willfully misleading ? Are you for real ?

How many time have I said in this thread that I don't have a full set-up with jag HW, thus can test only in VJ ?

 

How's posting results from VJ 'willfully misleading' ? I would expect better from you...

 

 

 

It's not just that they are not machine-exact, but that by their very nature they are inherently more misleading because they are claims made based on conditions that exist outside the bounds of a working game framework - a tiny loop of code abstracted from anything like working game conditions. Even if they were on real hardware they would be utterly meaningless because doing some arbitrary little non-test of one tiny fraction of what would be required within a working game environment is misleading.

Dude, no offense, but if you don't want to look absolutely ridiculous, you would do yourself a favour if you commented only on non-technical stuff - or let CyranoJ / Linko chime in.

 

You are basically stating that the whole 3d scene transformation and rendering is just a 'tiny loop of code' and that the Klax gameplay (!) will - somehow - be way more CPU-intensive than the 3d scene rendering.

 

That is , however, a nonsense in the context of the Klax gameplay.

 

The gameplay loop, aside from instructing DSP to play the music (how much CPU usage Linko ?), and checking the input, will only monitor an array of 25 bytes - e.g. upon placing a tile it will check for the currently-active combination of target tiles (and update score accordingly).

Note, that the score-update (a simple value increment ) and the target-combination check only happens during the frame when the tile is actually placed, not while they are falling down, at which point only the position of the 5-10 falling tiles is updated.

Then, every few seconds it will pick another tile from the pre-calculated array (at level load) and put it into array of actively rolling tiles. That's literally few instructions-worth of code.

 

That is not even close to 1% of the effort the CPU must undergo when transforming the vertices and rasterizing the polygons.

 

I cannot comment on the music, since I do not know how much CPU load / bandwidth will the DSP code playing the music consume. But I somehow doubt it would be more than 15-20% for a simple tune.

Either way, there is a resident audio expert on these forums, so he can perhaps chime in.

 

 

He entered this topic when GGN made a joke about making Klax in 3D at a very low frame rate:

 

It seems that he was attempting to prove that this claim of 10fps to be wrong, and by a long way, but he's not offering anything worthwhile to counter it.

 

He's only rocked up at AA since js2 closed. Prior to this he never hid his dislike for thiis forum and its users, in fact he openly relished a bout of AA-bashing... maybe because that kind of thing was fashionable there in the past. Who knows. Nobody cares.

 

All we do know is he's shown up here with a lot of bravado and is having the spotlight shined on his "beautiful" offerings for very little effort indeed. Bravo.

 

Yes, as I already said **multiple** times, I was trying to prove that 10 fps should be doable, even using C, without a single bit of ASM or GPU/DSP/Blitter cheating. Why is this so hard for you to finally acknowledge that ? I just took a short break (2 weekends) from my current PC project.

 

For me it was just a small challenge and I did improve my jaguar 3D routines, fixed few bugs and tested some new stuff. And it was fun !

 

 

Why on earth are you trying to talk down my whole effort - btw - you did not really do anything here except shedding negative light on my whole effort - is completely beyond me...

 

 

One more thing, I did not hear a single negative word from JagChris that he does not want his thread to contain my tech-blogging. If he said that he does not desire to see any temporary results, I would not post them.

 

So why are you so angry of my blog-style technical posts ? Just ignore them if you don't want to read them. Geeeeez......

[VladR]

Not arguing, merely pointing out the futility of the benchmarks...

If you feel it has to be repeated 10 more times, feel free to do so

 

 

but those 40 odd blocks (or was it 30) need to rotate down the ramp... and since your not using sprites you'll have to use rotational transformations (but you knew that, right?) - all those MULS won't slow it down at all, will they? (Just asking.......)

You haven't been paying attention

I said at the beginning, the tiles would not roll, but go/move/translate. The resolution is just too blocky for rotation of few pixels worth of cube (save for when they are closest).

 

I simply believe it will look better if they don't rotate, just move with the conveyor belt.

 

Plus, if the performance allows (which it looks like it will) a simple edge antialiasing would smooth out the results.

[+CyranoJ]

....Yeah.....

 

And sh3 is correct, above. Your results are misleading because your 'game code' is non-existant. What you have appeared to show is a bunch of plot, drawto commands. But still, if you want to ignore everything and carry on, go ahead. You'll be in for a shock when you find out how many cycles you lose for that stuff. Also, as I said, you need to vsync - nobody likes tearing, and this isn't a DirectX environment when people turn that stuff off to boost frame rates on their rigs.

 

BTW, I have benchmarked your HERO on real hardware, and yes, it gets 3fps. You might want to realize that not all Jaguars are created equal, and that your base libraries don't actually work properly, or you might end up with a VJ only game - which is very frustrating after you put in minutes of work on something

 

 

 

That is not even close to 1% of the effort the CPU must undergo when transforming the vertices and rasterizing the polygons.

 

I simply believe it will look better if they don't rotate, just move with the conveyor belt.

 

[VladR]

....Yeah.....

 

And sh3 is correct, above. Your results are misleading because your 'game code' is non-existant. What you have appeared to show is a bunch of plot, drawto commands. But still, if you want to ignore everything and carry on, go ahead. You'll be in for a shock when you find out how many cycles you lose for that stuff. Also, as I said, you need to vsync - nobody likes tearing, and this isn't a DirectX environment when people turn that stuff off to boost frame rates on their rigs.

 

BTW, I have benchmarked your HERO on real hardware, and yes, it gets 3fps. You might want to realize that not all Jaguars are created equal, and that your base libraries don't actually work properly, or you might end up with a VJ only game - which is very frustrating after you put in minutes of work on something

Which build of H.E.R.O did you get your hands on ?

 

I am pretty sure it wasn't this one:

[+Stephen]

I could buy the move to 3D, but non-rotating tiles? No way. You really need to play Klax to understand what made it a good game, as you are stripping everything that made it good out! Even on the little 6502 Lynx, there was a kick ass version of Klax. It featured animated and scaled graphics and digital audio which did not stop the gameplay. I wouldn't accept anything less on any system with something more than a 6502 in it. Sorry.

[BillyHW]

What's with all the hate on this thread? Some dude is in the very earliest stages of trying to program a neat 3D take on the most awesome puzzle game ever, Klax. Give him some time to figure things out. If it works it works, if it doesn't it doesn't. I'm interested in seeing how this progresses. And frankly, what's interesting to me is the 3D. There are a bazillian 2D version already out there. And I think the Jaguar is the perfect system to try this out on because clearly the earlier generation systems (SNES, Genesis) were not capable, and clearly the later-in-the generation systems (PS1, N64) were. So it's a formidable challenge on the Jaguar. If a 3D Klax had come out back in the day, I would have been very excited about it.

[VladR]

I could buy the move to 3D, but non-rotating tiles? No way. You really need to play Klax to understand what made it a good game, as you are stripping everything that made it good out!

I'm not sure how removing one visual fluff can be interpreted as 'stripping everything good out' ? So, the rotation of the pieces was the pinnacle of the game somehow ? That is interesting... Granted, I only watched videos. But my understanding was that player is supposed to catch&drop the colors in certain order to get the 'klax'...

 

Even on the little 6502 Lynx, there was a kick ass version of Klax. It featured animated and scaled graphics and digital audio which did not stop the gameplay. I wouldn't accept anything less on any system with something more than a 6502 in it. Sorry.

That comparison is invalid. Those were prerendered sprites. Since when do pre-rendered sprites equal real-time 3D graphics ? I mean, this has been the de facto primary technological problem and/or distinguishing factor between platforms for last 3 decades. And it still is a problem, since even modern gfx cards need to get faster to cope with the 3D rendering demands.

 

You do not have to be a 3D programmer to understand that it is completely unrealistic to expect identical level of quality in real-time 3D rasterizing compared to 2D pre-rendered stuff.

 

 

 

Now, that being said, I'm not saying jag could not handle, even just using C language, the rotation in real-time. It sure could. But I designed all the data structures and transformation and rasterizing algorithms around the tiles not being rotated. I'm certainly not going to redesign it for a questionable visual element.

 

But, I understand the sentiment - if you liked the game, you are correct in expecting a certain behaviour of the game. Sorry, I didn't view this feature as important - it was merely distracting in videos to me...

[VladR]

What's with all the hate on this thread? Some dude is in the very earliest stages of trying to program a neat 3D take on the most awesome puzzle game ever, Klax. Give him some time to figure things out. If it works it works, if it doesn't it doesn't. I'm interested in seeing how this progresses. And frankly, what's interesting to me is the 3D. There are a bazillian 2D version already out there. And I think the Jaguar is the perfect system to try this out on because clearly the earlier generation systems (SNES, Genesis) were not capable, and clearly the later-in-the generation systems (PS1, N64) were. So it's a formidable challenge on the Jaguar. If a 3D Klax had come out back in the day, I would have been very excited about it.

Oh my god. Dude ! What are you doing ? Quick, avoid the stampede ! Jump the bashing / Defenestration bandwaggon, while you can

 

So you must be the 'masses' I've been 'misleading' all this time

 

Anyone else ?

[VladR]

 

Note, that the score-update (a simple value increment ) and the target-combination check only happens during the frame when the tile is actually placed, not while they are falling down, at which point only the position of the 5-10 falling tiles is updated.

Then, every few seconds it will pick another tile from the pre-calculated array (at level load) and put it into array of actively rolling tiles. That's literally few instructions-worth of code.

So, I just went through the YT video of about 20 different ports of Klax here :

 

And I found out that all of the ports, upon getting the klax combo, actually pause the game completely ! They take their sweet time to flash the klax combination, update the score and some time after that (easily more than 50 frames), they let the tiles fall again and give the control back to player.

 

In other words, it is completely irrelevant how long does that particular piece of code take, since the 'klax' experience dictates that the game halts until the combo tiles are gone and score is updated.

 

 

So, this leaves the 'gameplay' portion to updating input, letting DSP know about the address of audio data and updating positions of animated tiles (simple increment) along with very infrequent (once every 50-100 frames) initialization of new cube from the pre-calculated array of cubes for current level.

[JagChris]

I think somewhere along the way a disagreement started and perhaps everyone needs to word things better or more polite. I dunno. But yeah these little proof o' concept things are rather frustrating. That 3D ramp looks really nice and its frustrating it will probably go no farther than that.

[JagChris]

Here Vlad, play some Klax

 

http://www.retronintendogames.com/k/klax/

[+Stephen]

What's with all the hate on this thread? Some dude is in the very earliest stages of trying to program a neat 3D take on the most awesome puzzle game ever, Klax. Give him some time to figure things out. If it works it works, if it doesn't it doesn't. I'm interested in seeing how this progresses. And frankly, what's interesting to me is the 3D. There are a bazillian 2D version already out there. And I think the Jaguar is the perfect system to try this out on because clearly the earlier generation systems (SNES, Genesis) were not capable, and clearly the later-in-the generation systems (PS1, N64) were. So it's a formidable challenge on the Jaguar. If a 3D Klax had come out back in the day, I would have been very excited about it.

There is no hatred whatsoever on my part. I have complimented Vlad many times on his coding skills, and I am quite impressed with his VRBasic. My comment strictly applies to a conversion of Klax. I happen to really like the game, played the shit out of it on my Lynx. I would not want a Jag port to give up the visuals and hectic gameplay of the original. Don't do 3D for the sake of doing it. It's the same trap as the Trameil era Jag. Texture map everything and drop from 30fps to 7fps because we need textures. No you don't. The animations and hectic sound effects really got you involved in the game. When you had 20+ tiles rotating towards yoy, and you could SEE and HEAR every one of them rotating, getting ready to fall off the edge, it made you nervous. Just watching something move towards me magically floating on a conveyor belt is NOT going to convey any sense of urgency.

 

You need to understand a game and what made it fun and difficult to play and replay, not just closely approximate the visuals of it. So yes, taking out the rotation of the tiles would be a huge disservice to the original game mechanics.

[BillyHW]

I agree with the sentiments on tile rotation, it just wouldn't be Klax without it and the accompanying sound effects, but how 'bout we just let him get it working without rotation and then someone can see if it can be added later. There's no Christmas shipping deadline to meet, and anyone else is free to try and program their own version.

Edited February 26, 2014 by BillyHW

[Rybags]

Remove the rotation?

 

May as well just program it as if it was one of those Nintendo Game & Watches from the early 80s.

[JagChris]

lol. I agree. Vlad you should really grab Klax for whatever system you own. They made 'greatest hits' packages for damn near every console. At least try it for a while. Seeing it play is one thing but actually playing it and getting into it will surely give you a way better feel for it.

 

I got it for my cousin for Christmas on his ps2. He left it on for 3 days at one point just to get to the 100th level.

[sh3-rg]

How many time have I said in this thread that I don't have a full set-up with jag HW, thus can test only in VJ ?

Nobody is questioning the use of VJ to create games, we do that. I questioned your metrics, their value and purpose. If you wish to offer meaningful performance claims, either sort out your Jaguar, post a binary or simply resist the urge.

 

How's posting results from VJ 'willfully misleading' ?

If you can't see why at this point, there's little hope you ever will.

 

Dude, no offense, but if you don't want to look absolutely ridiculous, you would do yourself a favour if you commented only on non-technical stuff - or let CyranoJ / Linko chime in.

None of your content is technical or beyond my comprehension. I might do art/design for Reboot but that's not the limit of my interests or understanding.

 

You are basically stating that the whole 3d scene transformation and rendering is just a 'tiny loop of code' and that the Klax gameplay (!) will - somehow - be way more CPU-intensive than the 3d scene rendering.

Attempting to straw man what I say is unproductive. Only comment on what I actually say and not what you wish I'd said in your eagerness to prove me wrong.

 

I never stated anything relating to portions of your code and how they might compare with others, especially ones that don't even exist yet. Why would I?

 

What I stated was that your "evidence" of very high "frame rates" under emulation is to all intents and purposes worthless. Offering them for public consumption is misleading and would only impress someone who didn't know any better.

 

This kind of stuff:

 

I turned off the double buffering to see how fast it really is... the cube-rendering part (including depth-based shading) runs in 100 fps

... is meaningless and you know why. I can't see any value in offering it other than to make anyone who doesn't understand say "WOW, that VladR sure is a great coder!" because anyone else is just going to shake their head and shrug. When you make such claims outside of not only the context of a working game (which as you well know is not just limited to "the "Klax gameplay (!)" - you're not even reading controller input and have no audio for a start.) but when you're not saving and restoring the screen underneath your generated objects and you make claims of "frames per second" but you're not actually generating frames or even syncing to the VBL, what is the point?

 

That's why I said it's willfully misleading. It's either that or you actually do believe it has merit, which itself would be more telling - take your pick.

 

The gameplay loop, aside from instructing DSP to play the music (how much CPU usage Linko ?)

Looks like you're in for some interesting times. If only audio replay was just a matter of clock cycles and it didn't impact the bus bandwidth in any way. Oh, btw, VJ's quite forgiving in that respect you'll be pleased to hear, so when you do finally get around to adding in audio to your "games", be sure to expect another order of magnitude in discrepancy between your emulated Jaguar and a real one.

 

This is why people have been telling you to think about what your are doing and to learn walk before you run. You simply have to learn more about the Jaguar and how it works or you will never reach the stage of producing and releasing a complete game, not one that works on real hardware at least. You are only setting yourself up for future disappointment and yet again we'll have had someone come along and make claims about their Jaguar coding prowess only to be left with nothing but a few tech demos, pages of forum posts and yet more disgruntled users. If all you are doing is coding exercises and tech tests, do it in the coding section away from the gamers and fans who might not understand that instead of hijacking threads on other subjects, or ask JagChris to ask a mod to move it there... save any more guys getting their hopes up needlessly.

 

That is not even close to 1% of the effort the CPU must undergo when transforming the vertices and rasterizing the polygons.

 

You do highlight here a lot of effort that you are going to in order to generate non-animated tiles in a 3D version of a 2D puzzle game that is already represented in 3D... we might need yet another thread to discuss that.

 

Everything you're doing is all the stuff you get "free" and better when you pre-generate depth-shaded, light-sourced, anti-aliased, beautiful sprites... just sayin'.

 

Yes, as I already said **multiple** times, I was trying to prove that 10 fps should be doable

And as GGN already said jokingly, if someone were to produce a 3D ossum Klax it'd likely run at 10fps. He didn't say it'd be impossible to run even at 10fps. Nobody required proof for such a low target to aim for as it was offered in jest being an unplayable, broken game - something the Jaguar's catalogue of 3D games isn't devoid of.

 

Regardless, the only way to prove that even a 10fps "3D" Klax was doable, would be to produce a complete game and run it on hardware - not just a small subset of a complete game devoid of everything that made it a game and then have it running under emulation.

 

Why is this so hard for you to finally acknowledge that ? I just took a short break (2 weekends) from my current PC project.

The length of time you have spent doing what you are doing is irrelevant, but if it somehow makes you feel better about things feel free to keep reminding us. It won't somehow make your benchmarking any less pointless. Plus, your posts here suggest you spent a few evenings at it in between those weekends, but like I said, irrelevant.

 

For me it was just a small challenge and I did improve my jaguar 3D routines, fixed few bugs and tested some new stuff. And it was fun !

Jolly good. But to prove or disprove the 10fps joke comment, you would have to go much further than you have so far and your measurements would have to be from real hardware in order to have any merit.

 

Why on earth are you trying to talk down my whole effort - btw - you did not really do anything here except shedding negative light on my whole effort - is completely beyond me...

That's not at all what happened. I questioned your metrics, their value and purpose.

 

So why are you so angry of my blog-style technical posts ? Just ignore them if you don't want to read them. Geeeeez......

I seem to remember you'd do this to Tursi on js2. Why do you insist on putting words into people's mouths?

 

I'll repeat...

 

That's not at all what happened. I questioned your metrics, their value and purpose.