pouët.net

Go to bottom

Assembly Summer oldskool compo rules

category: parties [glöplog]
branch: if you can provide a high quality recording, that would also probably be sufficient :)
added on the 2016-02-10 07:46:59 by britelite britelite
ok, if third option wins that's possible. but i prefer second option. cool
added on the 2016-02-10 08:32:59 by MuffinHop MuffinHop
Remember to vote no. 2, kids. Make oldskool compo great again! #DOS2016
added on the 2016-02-10 08:50:23 by noby noby
I think these choices are the wrong idea entirely. Demos should be judged on their own merits compared to the hardware they are running on. Have the oldskool compo include both #1 and #2!
added on the 2016-02-12 07:27:00 by trixter trixter
I think these choices are the wrong idea entirely. Demos should be judged on their own merits compared to the hardware they are running on. Have the oldskool compo include both #1 and #2; the audience is smart enough to figure out what is impressive and what isn't.
added on the 2016-02-12 07:29:43 by trixter trixter
Quote:
Demos should be judged on their own merits compared to the hardware they are running on

Yes, that would be ideal, but the truth is that the majority have no clue about most of the platforms the demos are running on
added on the 2016-02-12 07:30:44 by britelite britelite
especially at asm...
added on the 2016-02-12 10:47:39 by wysiwtf wysiwtf
At parties like Revision a lot of people don't have a realistic idea of the difference in performance between an a500 and a a1200 with 060. At Assembly even less.
added on the 2016-02-12 10:48:41 by dodke dodke
In a world where a sizeable amount of the attendees at *Revision* believe the AGA/060-stuff they see in the compos actually runs on vanilla amigas from '87 or '92. :D
Does oldschool platform need to be subdivided? :-o

Seriously though, a compo with similarly capable hardware would be more interesting than one that pits VIC-20 against 486. While there probably aren't enough entries to justify multiple oldschool categories, how about shifting the focus each year? Maybe coordinate with other big parties so that all categories get at least one "high profile" compo every year.

The category criteria are difficult though. Except for Amiga and Atari there aren't many releases for platforms between 8-bit and modern computers. Most of them are perhaps too exotic (X68000, FM Towns, PC-98).

By the way, isn't Pentium 133 a quite unfair advantage compared to 060 in the second option? I think 486 would be a more equal alternative.
added on the 2016-02-12 11:20:57 by absence absence
You're both underestimating and overestimating the Assembly audience. Underestimating because a demo on limited hardware is more than capable of winning the oldskool compo against more capable platforms - IF it's a good demo according to Assembly audience criteria. (See: Robotic Liberation.)

The oldskool demo compo at Assembly is never going to be an objective comparison of technical accomplishment, however much you fiddle with variables and platform lists and tilt the playing field.

*Any* compo at Assembly is evaluating one thing: your ability to entertain 5000 nerdy Finnish teenagers. All the rule tweaks in the world are not going to stop a demo featuring Nyan Cat and valkyrie warriors with big tits from winning against a moody radiosity-shaded flythrough of a cathedral, whatever platforms they're running on. That is the nature of the Assembly compo experience, and if you're not willing to accept that deal a demo maker, then to be honest you're probably better off submitting your demo somewhere like Forever or X instead. So why not embrace that? Allow a wide range of platforms, and give demo makers the freedom to choose whatever tools will allow them to make the most entertaining demo.
added on the 2016-02-12 11:23:29 by gasman gasman
Quote:
By the way, isn't Pentium 133 a quite unfair advantage compared to 060 in the second option?

Well, considering that some seem to be upset that they don't get a GUS in a 486 when competing against a C64 or A500, I'd say the Pentium 133 is probably too weak for them when competing against a 060-equipped Amiga/Atari ;)
added on the 2016-02-12 11:24:37 by britelite britelite
Quote:
The oldskool demo compo at Assembly is never going to be an objective comparison of technical accomplishment, however much you fiddle with variables and platform lists and tilt the playing field.

Well, to be honest, it's more about how to get more entries. The only time we actually got a good amount of entries in the past 10 years, was last year when the compo was in it's most limited form (Amiga only).

If we're to go in the "open for all" direction, why not just skip the oldschool compo altogether and have everything compete in the real wild?
added on the 2016-02-12 11:29:06 by britelite britelite
Quote:
Well, considering that some seem to be upset that they don't get a GUS in a 486 when competing against a C64 or A500, I'd say the Pentium 133 is probably too weak for them when competing against a 060-equipped Amiga/Atari ;)


Good point. :) 286/EGA/beeper vs C64 or 386sx/VGA/Adlib vs A500 sound like more interesting matches to me.
added on the 2016-02-12 11:30:47 by absence absence
Quote:
The oldskool demo compo at Assembly is never going to be an objective comparison of technical accomplishment


I'm guessing the purpose of this thread is more about convincing people to participate in a compo by catering to their whims and misconceptions, not about creating any sort of true fairness.
Obviously I'm all for this as well. :D
Argh, apparently I'm just repeating whatever some finn has previously said today..
Quote:

By the way, isn't Pentium 133 a quite unfair advantage compared to 060 in the second option? I think 486 would be a more equal alternative.

Wrong. 68060 shares a lot of architectural features with Pentium 1. Both have a superscalar in-order dual instruction pipeline config and an instruction decoder(CISC->RISC) . However 68060 FPU is not pipelined and is therefore it is slower in floating point performance but like stated in this thread before you should use fixed point anyways. Just clock both 060 and pentium at the same frequency and they are close to equal.

But anyways I'd be very happy with a fast 486, something like 66Mhz.
added on the 2016-02-12 14:24:19 by MuffinHop MuffinHop
Quote:
Just clock both 060 and pentium at the same frequency and they are close to equal.


Yeah, thing is, there aren't many 133 MHz 060s around. And why on earth wouldn't you use the FPU on Pentium to gain further advantage over the 060?
added on the 2016-02-12 14:52:56 by absence absence
What other reasons are there for using the FPU at runtime other than laziness? (in a program intended for mid 90s hardware and screen resolutions)
added on the 2016-02-12 15:15:55 by dodke dodke
@absence: First Pentium was running at 60 or 66 MHz.
added on the 2016-02-12 15:57:20 by noname noname
Quote:
What other reasons are there for using the FPU at runtime other than laziness? (in a program intended for mid 90s hardware and screen resolutions)

I think you have it backwards. Not using the Pentium FPU would be laziness, as it requires quite a bit of work to schedule the instructions properly. When you get it right however, you can have the CPU execute integer and floating point instructions simultaneously which needless to say is a performance advantage when used creatively. There's a reason Quake (mid-90s hardware and screen resolutions as you say) used floating point arithmetic instead of fixed point, and I doubt it was John Carmack being lazy.
added on the 2016-02-12 16:04:07 by absence absence
Quote:
@absence: First Pentium was running at 60 or 66 MHz.

We're discussing the Pentium 133 in the second option suggested in the first post of this thread, here's what i wrote:
Quote:
By the way, isn't Pentium 133 a quite unfair advantage compared to 060 in the second option?
added on the 2016-02-12 16:08:15 by absence absence
I kinda do agree with Dodke, but on Pentium some floating point operations can be paired and some fixed point can not be.

Here's something I hope is possible(=compo crew can get the hardware needed) and fair for option 2:

PC - 486DX4 at 75Mhz, with ISA regular VGA and GUS
Amiga AGA - 060 at 66Mhz

486DX4 at 75Mhz does not superscalar but has the little boost against 060 superscalar at 66Mhz. And not having Super VGA means that most realtime graphics look roughly same on Amiga and PC except Amiga can do HAM-graphics which creates a bit of fair difference between the two when PC has GUS that is better than Paula. Also having ISA VGA card on PC means that bus speed is closer to Amiga and C2P conversion. I prefer this because then you have a bit of fair difference (in taste?) but still fairly close platforms in performance.

Left Falcon out because I'm not knowledgeable enough to dare say anything about it.
added on the 2016-02-12 16:13:09 by MuffinHop MuffinHop
Laziness is an important factor in any compo, IMO. Micron is saying basically the same thing. Every added obstacle reduces motivation and the number of entries. The same with many other things. "Real wild" lowers motivation, because you have to assume that it will be even harder for the audience to judge the entries, and be interested in it. The audience has to have a good idea of what they're being shown. (though not neessarily the correct idea)

It's about magic, tricks, creating illusions of something impossible or cool or something. You cannot create that magic touch if the audience doesn't have some kind of sense of what is normal on that platform. If you take this away, demos become pointless, IMO. In other words, no motivation, no entries. If you only want to create cool visuals, make a video.

So, cost/benefit analysis.
Cost = time and money. How difficult will it be to code and test, do you have to buy hardware, set up new and unfamiliar build systems, study new techniques.
Benefit = how rewarding will it be to get the prod shown, will the audience think it's cool, will you get kicks from making it.

The compo rules affect both sides of the equation. Allow all platforms, and benefit goes to zero. Allow only very hard-to-get and hard-to-code-for platforms, and cost goes to infinity.


By the way, is anyone interested in testing a 60 Hz VGA mode 13h? I've made a set of CRTC register settings that creates a 60 Hz 320x200 mode without doing any "un-chain-four" stuff, but I don't have a CRT monitor to test it with. You have to be a bit adventurous because messing around with CRTC settings might break your monitor.
added on the 2016-02-12 16:22:45 by yzi yzi
@yzi: Test with an LCD monitor.
added on the 2016-02-15 09:57:20 by trixter trixter

login

Go to top