Boring Friday night in? Looking for some hot, esoteric programming challenges? Give my "Low-Rez Shadertoy: The Game" a go!

Write shaders in some made up assembly language, match your target, win prizes*

This is an alpha version and I'm looking for feedback. Please boost/share, thanks!

http://lrst-alpha.s3-website.eu-north-1.amazonaws.com/

LRST

@evilpaul_atebit I tried to zoom in on my phone but it resized in a way that actually zoomed out. I guess I'm gonna look on my laptop later today instead.
@endrift Huh.. interesting. As you need keyboard/mouse, it never occurred to me to try it on iPhone but that zoom behaviour is very weird indeed :/
@evilpaul_atebit Ok finally tried it on my laptop in Firefox and, uh, is there supposed to be text? I'm not seeing any text. I have no idea what I'm doing without text.

@endrift Ah yes.. someone else reported that. Do you have Firefox in a stricter privacy mode than default? I think the way I load the font makes Firefox think that I'm trying to fingerprint the user. I'm not and the game doesn't try to track you or send any user data anywhere. If you disable strict checking for this site then it should work.

I will try to find a different way to load the font that doesn't upset Firefox.

@endrift I just deployed a new version that may fix this..
@evilpaul_atebit Yup, it works now! This is pretty neat. I have a 100% match solution to #3 in only 2 instructions, which is more than that 90% goal
@endrift Ahh yes.. there is a two instruction solution for that one - it really should be the target. I'll fix that. Thanks! :)
@evilpaul_atebit I also have a 100% solution for #4 in 3 instructions; it works basically the same way
@evilpaul_atebit is it obvious that I've written shaders in assembly before? 😉​
@endrift Is that "stage 1-4: Flag"? The current target solution for that is 3 instructions:
lsr r1,#5,r1
mov.nz #6,r0
mov.z #1,r0
@evilpaul_atebit it is but that's not how I do it. That's probably a better way and how I ended up doing the diamond ones. I did a mulu + add
@endrift I’m really excited that people are coming up with different solutions that do the same thing un the same number of instructions. I’m not surprised (I’ve done enough asm size coding to have expected this) but it’s exciting to see it happening :)
@evilpaul_atebit I dunno what your intent for this puzzle was, but uh, I managed a lot less than your target
@endrift I’m not near the code to check… but that looks like the “correct” solution. In the version of the code that’s deployed, those target numbers are typed in by hand and I messed a few of them up - setting the target too high. In the current (unpublished) version the counts are derived directly from the code that produces the target, so will be more accurate. Also in the current version are SIMD extensions, which will enable shaving three more instructions off that solution ;)
@evilpaul_atebit oh man I was craving SIMD for the one before this. That's so good to hear