Mobile demo compos
category: general [glöplog]
whizzter: i totally disagree about pc hw (and btw agenda is 4 years old now). but youre right in a way that whether to merge mobile and pc does depend on what pc hardware youre talking about.
to me, the PC target is "the best desktop PC you can build for demos, for about 1000e without a screen" - a definition also used by some demo parties. that pc would probably contain the latest greatest GPU you can get for around 300-400e, and would give more than 10x the performance of a new low end laptop's integrated GPU. we are not talking about supercomputers with multiple 10,000e GPUs here - we're talking about a PC that a regular person could potentially buy today for a reasonable amount of money if they wanted to.
the problem with PC is that there's such a huge range of levels of hardware. netbooks and most laptops are obviously at the lower end, but people dont upgrade their pc that much anymore either. so what do you aim for? if i target e.g. a geforce 680 - a high end card in many people's eyes, but 2 years old - its going to give half the power as if i targeted say a radeon 290. but by targeting either id end up with a demo that hardly anyone can run realtime at a good framerate - because most people are sitting on even older GPUs, or on a laptop. for reference, by macbook pro 15 which is about 2 years old and has an ati 6870 in it runs about 1/10 the performance in many of my test cases as my home shuttle pc which cost £500 a few months ago.
if i choose to target below the high end, it directly limits what i can do and the demo i can produce. something that runs at 10fps at best is not practical to use. something that runs at 20-30fps is potentially practical. yet it seems somewhat pointless to target a bit below the high end because people still wont be able to run it.
the only worthwhile thing would be to make something that runs on a netbook - but then the results are so far away from what can be done on a high end pc that it doesn't feel like the same compo.
in summary: a mobile device is not a real pc, even if it runs gl.
maybe you could have a high end hw compo (i.e. pc) and a low end hardware compo (netbook, mobile device), if there were the entries to support it. hey, a netbook democompo might feel better for some pc coders than the high end.
also,
this is more of a sad reflection on the state of pc democoders than the hardware itself.
to me, the PC target is "the best desktop PC you can build for demos, for about 1000e without a screen" - a definition also used by some demo parties. that pc would probably contain the latest greatest GPU you can get for around 300-400e, and would give more than 10x the performance of a new low end laptop's integrated GPU. we are not talking about supercomputers with multiple 10,000e GPUs here - we're talking about a PC that a regular person could potentially buy today for a reasonable amount of money if they wanted to.
the problem with PC is that there's such a huge range of levels of hardware. netbooks and most laptops are obviously at the lower end, but people dont upgrade their pc that much anymore either. so what do you aim for? if i target e.g. a geforce 680 - a high end card in many people's eyes, but 2 years old - its going to give half the power as if i targeted say a radeon 290. but by targeting either id end up with a demo that hardly anyone can run realtime at a good framerate - because most people are sitting on even older GPUs, or on a laptop. for reference, by macbook pro 15 which is about 2 years old and has an ati 6870 in it runs about 1/10 the performance in many of my test cases as my home shuttle pc which cost £500 a few months ago.
if i choose to target below the high end, it directly limits what i can do and the demo i can produce. something that runs at 10fps at best is not practical to use. something that runs at 20-30fps is potentially practical. yet it seems somewhat pointless to target a bit below the high end because people still wont be able to run it.
the only worthwhile thing would be to make something that runs on a netbook - but then the results are so far away from what can be done on a high end pc that it doesn't feel like the same compo.
in summary: a mobile device is not a real pc, even if it runs gl.
maybe you could have a high end hw compo (i.e. pc) and a low end hardware compo (netbook, mobile device), if there were the entries to support it. hey, a netbook democompo might feel better for some pc coders than the high end.
also,
Quote:
Sure raymarching has pushed up the base requirements and you can always spawn more particles,etc but in terms of most artistically created content (models, images,etc) mobiles can run basically the same things as a PC.
this is more of a sad reflection on the state of pc democoders than the hardware itself.
Note: This is less a post of dissension than of nuanced agreement.
Maybe a range of demosceners have different valid goals *which can coexist*, and mobile can work for some of them.
-There's the traditional "peg the most badass hardware at the top of its performance curve".
-There's "communicate this idea".
-There's "rise to this particular visual or technical design challenge".
-And in the shallower end of the pool, there's "just make something".
A fictional story I read about vaudeville had a real few quotations in it which seem diagonally apropros here.
"Kid, go out and be bad. Come back and see me when you get good, and I'll help you all I can." -- Al Shean
"Now that vaudeville is dead, there's nowhere for kids to go and be bad anymore."
--George Burns
The two fictional characters discussing these quotations have an interchange right after the Burns quote.
M: "What about the Fox Network?"
W: "You got him there."
Yes, the people who are capable of making technically impressive and artistic demos should produce, and keep developing their conceptual and technical skills.
But the pipeline of new blood needs to keep pumping too.
There are a *lot* of new mobile developers out there now who are potential sceners, who may be less daunted by that platform than PC. It is also a platform that is more accessible, costwise, and more broadly distributed around the world than really good PCs (speaking as someone whose computer of choice gets hot enough to fry eggs when you foolishly try to play Minecraft or that new Shadowrun game).
The mobile world as noted also has marketplaces that could lead to more visiblity for the scene.
Let's not go the way of vaudeville in the age of movies.
Maybe a range of demosceners have different valid goals *which can coexist*, and mobile can work for some of them.
-There's the traditional "peg the most badass hardware at the top of its performance curve".
-There's "communicate this idea".
-There's "rise to this particular visual or technical design challenge".
-And in the shallower end of the pool, there's "just make something".
A fictional story I read about vaudeville had a real few quotations in it which seem diagonally apropros here.
"Kid, go out and be bad. Come back and see me when you get good, and I'll help you all I can." -- Al Shean
"Now that vaudeville is dead, there's nowhere for kids to go and be bad anymore."
--George Burns
The two fictional characters discussing these quotations have an interchange right after the Burns quote.
M: "What about the Fox Network?"
W: "You got him there."
Yes, the people who are capable of making technically impressive and artistic demos should produce, and keep developing their conceptual and technical skills.
But the pipeline of new blood needs to keep pumping too.
There are a *lot* of new mobile developers out there now who are potential sceners, who may be less daunted by that platform than PC. It is also a platform that is more accessible, costwise, and more broadly distributed around the world than really good PCs (speaking as someone whose computer of choice gets hot enough to fry eggs when you foolishly try to play Minecraft or that new Shadowrun game).
The mobile world as noted also has marketplaces that could lead to more visiblity for the scene.
Let's not go the way of vaudeville in the age of movies.
oh well.. power in numbers, let's see some more mobile demos first. let's hope nvscene's shield compo is a good inspiration for some ;) there's no visibility when there's only a handful of mobile demos available (which is how it is). and with visibility hopefully comes interest from outer circles.
Quote:
Maybe a range of demosceners have different valid goals *which can coexist*, and mobile can work for some of them.
-There's the traditional "peg the most badass hardware at the top of its performance curve".
-There's "communicate this idea".
-There's "rise to this particular visual or technical design challenge".
-And in the shallower end of the pool, there's "just make something".
I guess that makes a good argument for allowing mobile productions in the PC compo. 3 out of 4 points are not about the performance/feature level of the platform.
Judging by the few Poueters who still run demos on their computers, my iPhone has a better GPU than they seem to have, so I say stick the mobile demos in the "combined demo compo" - that's what it's named for anyway.
Side-argument: we've long passed the point where actual visual fidelity is any guarantee for success in a compo. Humor, concept and execution of simpler ideas are the prevailing elements here. Insisting that any single compo should be as homogenous as possible in the CPU/GPU-power-department is stupid and out-of-touch.
..especially since "everyone" watches demos on YouTube anyway.
Side-argument: we've long passed the point where actual visual fidelity is any guarantee for success in a compo. Humor, concept and execution of simpler ideas are the prevailing elements here. Insisting that any single compo should be as homogenous as possible in the CPU/GPU-power-department is stupid and out-of-touch.
..especially since "everyone" watches demos on YouTube anyway.
Quote:
I guess that makes a good argument for allowing mobile productions in the PC compo. 3 out of 4 points are not about the performance/feature level of the platform.
and surely by extension for lumping in amiga, c64, etc etc. into that one compo. and 64k and 4k for that matter.
and pixel, graphics, music too.. it's all art last time i looked at it!
My stuff is technically simple and doesn't require much from hardware (both because I don't really find that stuff interesting to do and because I'm coding it on an old laptop that can't really handle much more). Should I put my demos on the the hypothetical low-end compo then, to leave the other one for the few real sceners among us?
Word.
Quote:
Humor, concept and execution of simpler ideas are the prevailing elements here. Insisting that any single compo should be as homogenous as possible in the CPU/GPU-power-department is stupid and out-of-touch.
Word.
8bit demos are more interesting and impressive nowadays.
Quote:
to leave the other one for the few real sceners among us?
erm, who attached any value or quality to the low or high end? both are equally interesting, they just run on different hardware.
we could also just exclude fairlight from competition and then the artistic and technical feats of mobile and pc demos are pretty much interchangable thus mergable as a compo! ;)
This discussion reminds me of a comment I got for a software rendering intro I built some years ago with following statement:
And I just chose a mid-end PC without hardware acceleration which is imho also equally interesting.
Quote:
It's an odd and rather self-limiting choice of tools, that is all.
And I just chose a mid-end PC without hardware acceleration which is imho also equally interesting.
Quote:
and surely by extension for lumping in amiga, c64, etc etc. into that one compo. and 64k and 4k for that matter.
Amiga/PC has happened before. Amigas even won in such compos.
Such as Lapsuus.
Or Starstruck.
Mobile is closer to PC than Amiga is, so I don't see why not.
Quote:
16bit : amiga, atari and comparable
8bit : c64,nes and comparable
Where does 8088 fit? (16-bit registers, 8-bit bus, 4 cycles per byte access which IIRC is worse than 6502)
Quote:
Where does 8088 fit? (16-bit registers, 8-bit bus, 4 cycles per byte access which IIRC is worse than 6502)
I'd say demo-wise the 8088 is most comparable to 8-bit machines.
The 16-bitters tend to have more capable CPUs, graphics and audio hardware than an 8088-based PC (even if it's a fully equipped 9.54 MHz 8088 with VGA and SB, it still would be closer to a good 8-bit machine (think Batman Forever) than an Atari ST or Amiga).
You can't put Amiga and PC in the same compo!
That's not fair to the people who make the PC demos.
(:
That's not fair to the people who make the PC demos.
(:
Also, although oldschool and PC should be different compos, every now and then you get a Batman Forever which could wipe the floor with many PC demos.
In my mind its kinda modern platforms vs pre-1992 platforms. Mobile is generally modern. (I say generally because you just know some crazed masochist could come out of the woodwork without warning and make a tiny black and white demo on a Cityman 150 or a Polish guy appears with a streaming demo using NMT-Text. It could happen).
In my mind its kinda modern platforms vs pre-1992 platforms. Mobile is generally modern. (I say generally because you just know some crazed masochist could come out of the woodwork without warning and make a tiny black and white demo on a Cityman 150 or a Polish guy appears with a streaming demo using NMT-Text. It could happen).
ah shoot. you spoiled our competitive edge. the only way to win from fairlight is to incorporate as many batmans as possible. hm, how about a CarPlay demo about dwarfism then...
maali: dont worry about us, we're pretty much retired now anyway :)
*retarded
smash: How do you mean that it is a sad reflection of pc democoders?
What i'm talking in that sentence is that mobile hardware has enough power to render characters and other specifically created models people would have time to do in their spare time for demos, in the same sense as that the biggest games today don't have 30 people working on graphics like at the end of the 90s but rather hundreds (thousands?). Sure you'd be combining it with generated worlds but at that time you'd have the ability to select the fidelity level. (Basically like scali and metoikos mentioned as those 3 out of 4 points)
But if you have a good counter example about i'd love to hear it :)
As for the difference in targeting enthusiast GPU's and doing what's possible on that vs working integrated GPU targets, we've really had that problem since the day we left software rendering but as you also mention, time and hw evolution used to close the performance gap until recently when machines started to become good enough for people to not upgrade constantly.
Personally, since i haven't had a stationary PC in 14 years now and my laptops has been weighing 1-1.7 kg's for 10 years of those so i've been mostly youtubing demos for a long time and mostly pulled up demos to see what level my new machines was capable of.
The democoding i've done in that time has never been about pushing rendering (since i didn't ever own the hardware) but rather trying to squeeze content for size or just plain partyhacking to get something done and i'm pretty sure i'm not the only person with that focus. (Although i've started looking at doing coding for older platforms because it could be interesting to learn it to see if one could push hw there)
What i'm talking in that sentence is that mobile hardware has enough power to render characters and other specifically created models people would have time to do in their spare time for demos, in the same sense as that the biggest games today don't have 30 people working on graphics like at the end of the 90s but rather hundreds (thousands?). Sure you'd be combining it with generated worlds but at that time you'd have the ability to select the fidelity level. (Basically like scali and metoikos mentioned as those 3 out of 4 points)
But if you have a good counter example about i'd love to hear it :)
As for the difference in targeting enthusiast GPU's and doing what's possible on that vs working integrated GPU targets, we've really had that problem since the day we left software rendering but as you also mention, time and hw evolution used to close the performance gap until recently when machines started to become good enough for people to not upgrade constantly.
Personally, since i haven't had a stationary PC in 14 years now and my laptops has been weighing 1-1.7 kg's for 10 years of those so i've been mostly youtubing demos for a long time and mostly pulled up demos to see what level my new machines was capable of.
The democoding i've done in that time has never been about pushing rendering (since i didn't ever own the hardware) but rather trying to squeeze content for size or just plain partyhacking to get something done and i'm pretty sure i'm not the only person with that focus. (Although i've started looking at doing coding for older platforms because it could be interesting to learn it to see if one could push hw there)
they do not, however, have the power to render models that weren't created specifically.
Quote:
is it better for mobile demos to be in the "combined demos"-compo or the "alternative platform"-compo? Because AFAIK they're not really still being tossed in with "alternative platform" (atleast not outside of Finland :) - they either have their own compo (like we do at NVScene in a few weeks), or they're shown next to PC demos and web demos in the "regular demos"-category.
I wouldn't mind competing with PC demos, which would be better place for mobile demos, than the "toaster with self soldered LED screen" compo.
As for the "mobile devices are as powerful as PCs cause they run OpenGL"... Well not really. It comes down to the fact that mobile chips can use only a couple of watts of power, where as modern desktop GPUs are already knocking the 300 watt door as we speak. But yeah, OpenGL ES makes it really easy to make the move from PC to mobile. I recommend giving it a shot ;)
Quote:
Side-argument: we've long passed the point where actual visual fidelity is any guarantee for success in a compo. Humor, concept and execution of simpler ideas are the prevailing elements here. Insisting that any single compo should be as homogenous as possible in the CPU/GPU-power-department is stupid and out-of-touch.
Agreed.
I'd like to see more mobile demos (and ideally make some, but then i'd like enough time to finish what i'm working on for desktop first :)
Couple of points I can add here:
1. PCs aren't selling less because people aren't upgrading them as often (although that's probably part of it). They're getting used less because people are using other (i.e. mobile) platforms instead. There's a ton of evidence for that.
Wikipedia's stats got linked up on twitter this morning, I took a look. Feb 2013: 217m visits, 124m from windows. Feb 2014: 247m visits, 110m from windows. More visitors, yet less PCs.
Looking at the mobile stats: 2013: 67m, 2014: 110m. Note that equals the PC visitors now.
Also relevant: ~3x more visits from iOS devices than android. There's a lot more android devices out there, but a lot of them aren't used for much more than phone calls. Worth remembering if you want a big audience.
Stats are here btw: http://stats.wikimedia.org/archive/squid_reports/2014-02/SquidReportOperatingSystems.htm (android is bundled with linux, but there is a breakdown).
2. mobiles aren't *that* weak in performance. Take the current high end (I'm sure there's a lot more iPhone 5s' out in the wild than high end PCs :) and you have something roughly like a 2.5ghz core 2. I've not seen a mobile/desktop GPU comparison anywhere, but I'd guess mid-range from 5 years back maybe?
3. OpenGL ES 3 brings the API level up a fair way.
4. Some 'official' system for iOS demos would be good. The app store rules can be a problem though, apple reject stuff that doesn't have a useful purpose and a battery draining non-interactive thing that runs for 5 minutes then quits isn't all that useful :D (And I've heard of them rejecting a submitted demo because "demos aren't allowed, consider making a lite version instead" ffs!)
I've been thinking about that, and i reckon there are 2 ways to handle it. 1: put some interactivity in. 2: create a 'launcher' app with a nice menu that lets you choose a demo and watch it.
Couple of points I can add here:
1. PCs aren't selling less because people aren't upgrading them as often (although that's probably part of it). They're getting used less because people are using other (i.e. mobile) platforms instead. There's a ton of evidence for that.
Wikipedia's stats got linked up on twitter this morning, I took a look. Feb 2013: 217m visits, 124m from windows. Feb 2014: 247m visits, 110m from windows. More visitors, yet less PCs.
Looking at the mobile stats: 2013: 67m, 2014: 110m. Note that equals the PC visitors now.
Also relevant: ~3x more visits from iOS devices than android. There's a lot more android devices out there, but a lot of them aren't used for much more than phone calls. Worth remembering if you want a big audience.
Stats are here btw: http://stats.wikimedia.org/archive/squid_reports/2014-02/SquidReportOperatingSystems.htm (android is bundled with linux, but there is a breakdown).
2. mobiles aren't *that* weak in performance. Take the current high end (I'm sure there's a lot more iPhone 5s' out in the wild than high end PCs :) and you have something roughly like a 2.5ghz core 2. I've not seen a mobile/desktop GPU comparison anywhere, but I'd guess mid-range from 5 years back maybe?
3. OpenGL ES 3 brings the API level up a fair way.
4. Some 'official' system for iOS demos would be good. The app store rules can be a problem though, apple reject stuff that doesn't have a useful purpose and a battery draining non-interactive thing that runs for 5 minutes then quits isn't all that useful :D (And I've heard of them rejecting a submitted demo because "demos aren't allowed, consider making a lite version instead" ffs!)
I've been thinking about that, and i reckon there are 2 ways to handle it. 1: put some interactivity in. 2: create a 'launcher' app with a nice menu that lets you choose a demo and watch it.
oh yeah, and mixed pc/mobile makes sense to me, unless there's going to be enough mobile demos for a separate compo.
A high end phone is faster than some of the PCs people make their demos on, so it makes zero sense putting them in any other compo. And I can't see how the most common platforms in the world belong in the wild compo ;)
A high end phone is faster than some of the PCs people make their demos on, so it makes zero sense putting them in any other compo. And I can't see how the most common platforms in the world belong in the wild compo ;)