[HN Gopher] WebGL Fluid Experiment (2014)
___________________________________________________________________
WebGL Fluid Experiment (2014)
Author : maxraz
Score : 182 points
Date : 2020-08-05 11:57 UTC (11 hours ago)
(HTM) web link (haxiomic.github.io)
(TXT) w3m dump (haxiomic.github.io)
| haxiomic wrote:
| Hey, I feel bad for not updating this demo in 6 years! Back then
| it wouldn't run on iOS, today it could run with no trouble
|
| For people on iOS, here's another version with rainbow colors and
| multi-touch https://haxiomic.github.io/demos/rainbow-
| fluid/index.html
|
| If you like this sort of thing and want to learn more about it I
| recommend checking out https://www.shadertoy.com/ for more demos
| and https://thebookofshaders.com/ for tutorials
|
| And if GPU simulations is your thing then I think you'll enjoy
| what these two amazing guys are doing with shaders:
|
| - https://www.shadertoy.com/user/michael0884
|
| - https://www.shadertoy.com/user/wyatt
| seesawtron wrote:
| This is beautiful. Can you say how this is different from [0]
| if at all in terms of implementation?
|
| [0] https://paveldogreat.github.io/WebGL-Fluid-Simulation/
| haxiomic wrote:
| Similar implementation from a physics perspective - Pavel
| uses the same technique for simulating Navier-Stokes (Jo Stam
| Real-Time Fluid Dynamics for Games [0]). Pavel's has an extra
| step to enhance vorticity which gives you those higher detail
| swirls and uses a slightly different approach for input
|
| The big difference is in the rendering however, Pavel has
| added light shafts and bloom which are really neat effects.
| Additionally, he actually got round to publishing a native
| app which I think has been great for him!
|
| [0] https://pdfs.semanticscholar.org/847f/819a4ea14bd789aca8b
| c88...
| nitrogen wrote:
| FWIW the rainbow link gives this console error in Firefox on a
| Surface Book with Intel graphics: Uncaught
| TypeError: rgHalfFloatLinear is null
|
| I assume that means it doesn't support 16-bit float pixel
| formats; maybe this is something that could be detected and
| shown as an error message in the page?
|
| The original HN link works great, though, very cool!
| haxiomic wrote:
| Ahh that's a shame, yeah looks like 16-bit float textures
| aren't supported but 32 bit are. Thanks for the heads up - I
| should really be checking if either are supported here!
| monkeydust wrote:
| Therapeutic... I mean really.
| blauditore wrote:
| Tangentially related: For a long time I've always been looking
| for some software for particle simulations that is free or even
| open source, easy to get started, and at least somewhat visual.
| I'm not doing anything like that professionally, just want to
| play around with a virtual wind tunnel for the fun of it. I'm
| thinking of something like Blender where I would click together a
| 3d mesh, "turn on" wind and see what happens.
|
| Does anyone have tips or hints? Are my expectations perhaps
| unrealistic? I looked into OpenFOAM before but always eventually
| gave up, as it seems quite heavy on headless, scripted
| simulation, and would need some significant run-up time.
| eric_t wrote:
| Associate professor in fluid dynamics here. There is typically
| a split between the people who want good looking simulations
| for movies/games, and those who need accurate results for
| engineering. For engineering, you need more effort on how you
| resolve the complex geometry and how you model the turbulence
| in the flow. This adds significant complexity both for problem
| setup and solution time. The only product I'm aware of that is
| built for real-time, 3D flow design is Ansys Discovery:
| https://www.ansys.com/products/3d-design/ansys-discovery
|
| OpenFOAM is high quality, but as you've seen complex to use.
| There is a web-based GUI that can lighten the burden to get
| started somewhat: https://www.simscale.com/ I think they give
| you ~3000 simulation hours for a trial.
| vosper wrote:
| I've been trying to figure out where I should place a shelter
| belt on my property, as I've heard that the wind will rise up
| over the belt and then "dump" downwards. I don't want to
| position the belt so that the wind dumps on my house. I think
| it's essentially a fluid dynamics problem (I say this as
| decidedly not an engineer).
|
| I've only looked half-heartedly, but found nothing that can
| help me.
| jpm48 wrote:
| Specifically for fluid (and similar) you could have a play with
| mantaflow http://mantaflow.com/ it's relatively easy to build
| (you will need Qt and Python). Houdini (free apprentice version
| will work for fun) is great for particle effects and this sort
| of stuff as well https://www.sidefx.com/
| jchw wrote:
| There is also this one, and the corresponding iOS app (which I
| have.) It seems to run a bit faster on my phone.
|
| https://paveldogreat.github.io/WebGL-Fluid-Simulation/
| jedimastert wrote:
| The README links to the original as a reference!
| utf_8x wrote:
| Damn, the performance on that is really impressive.
| imvetri wrote:
| ha ha ha, outer space is a simulated fluid.
| imvetri wrote:
| with that assumption, it gets easier to build an eye observing
| an eye....observing an eye observing a thing.
| imvetri wrote:
| Neural network learning models are good, but it scales well
| only when NN watches a NN watching an NN being trained with a
| data.
| nullc wrote:
| To be sure that you're not a NN yourself, I should require
| you to draw a triangle.
| exikyut wrote:
| _[Triangle is distorted by unseen fluid]_
| inetsee wrote:
| The title seems to imply that it needs a touch screen device. It
| seems to work just fine on my desktop with a mouse.
| suyash wrote:
| beautifully made!
| designium wrote:
| That's very fun and relaxing!
| holoduke wrote:
| just out of curiousity, but how does the algorithm roughly works.
| does every pixel/particle position gets calculated? or is it
| optimized by smart grouping? is the GPU used for paralell
| computing? very interesting stuff to heavily optimize I guess.
| dang wrote:
| Discussed at the time:
| https://news.ycombinator.com/item?id=8502477
|
| also https://news.ycombinator.com/item?id=8325700
| madjam002 wrote:
| Nice, the effect reminds me of Powder Game (https://dan-
| ball.jp/en/javagame/dust/)
| otterpop wrote:
| At least give me some warning before you link to it; I'm gonna
| be playing it all day today :)
| ljf wrote:
| Man there was some game similar to this called Sand Sand
| Sand, that I was totally addicted to. I showed my kid
| recently and he loved it too, timeless fun!
| thedrbrian wrote:
| There's also the stand alone Powder Toy which has more
| powders and saves and such
|
| https://powdertoy.co.uk/
| snazz wrote:
| And the modern web-based incarnation Sandspiel, which was
| extremely popular on HN a while ago:
| https://news.ycombinator.com/item?id=18696291
| hansoolo wrote:
| Awesome!
| kipple wrote:
| At first I expected this to be some new version, but nope, it's
| the original. I love that this repo[0] hasn't been touched
| significantly in 6 years and it's still going strong.
|
| [0] https://github.com/haxiomic/GPU-Fluid-Experiments
| soylentgraham wrote:
| Given that the original commit is clearly a giant copy&pasta
| from somewhere else, not surprised there is little development
| or changes as time goes on
| soylentgraham wrote:
| Oh, my bad, this the copy&pasta one
| https://paveldogreat.github.io/WebGL-Fluid-Simulation/
| jchw wrote:
| Trying to figure out what you are suggesting with the
| copy/paste stuff. I looked at the first few commits in both
| the PavelDoGreat repo and the haxiomic repo and see no
| obvious signs of copypaste. PavelDoGreat's initial commits
| start with initializing WebGL, seems reasonable. haxiomic's
| starts with a README and then a code drop, again, doesn't
| seem too unexpected. I also doubt one took from the other
| since one is Haxe and the other is JS.
| monkeydust wrote:
| That's the one I was thinking of, thank you.
| peacefulhat wrote:
| iOS not supported. Ridiculous that iOS doesn't support WebGL 2.0
| ry4nolson wrote:
| https://news.ycombinator.com/item?id=24062471
| Razengan wrote:
| Would be even more fun if such things (including the Powder Game)
| were also affected by device
| orientation/gyroscope/accelerometers.
| nemacol wrote:
| This is fantastic. Runs really smooth, the color and design looks
| wonderful. I could lose few afternoons playing with this thing.
|
| Looking forward to the app. Would be fun to manipulate the sim by
| moving my iPad around.
|
| Potentially fun ideas...
|
| Allow me to place shapes/blocks down and lock them in place to
| see how the fluid moves around them.
|
| Let me put a small 'pump' in the liquid to keep constant or
| pulsed movement. add a couple dials for pump speed, etc. I want
| to put two pumps facing one another and watch the interaction
| where they meet.
|
| Different pool shapes - round, star, sphere?
|
| Allow for particles to reach the edge of the pool to fall off and
| then let me drop new ones into flowing water.
|
| Loads of fun. Thank you!
| _ZeD_ wrote:
| It reminds me of http://minimal.be/lab/fluGL/index80000.html
| codetrotter wrote:
| Doesn't support iOS
| lawlessone wrote:
| Or does iOS not support it?
| gen3 wrote:
| Here's one that works on iOS. It also has an app.
| https://paveldogreat.github.io/WebGL-Fluid-Simulation/
| Ataraxy wrote:
| Very satisfying effect.
| octernion wrote:
| i'm constantly amazed at what can be done with webgl these day
| (shame about the iOS support). surprised we don't see more
| immersive games with this tech!
| LoSboccacc wrote:
| all these simulators out there and still not one remake of plasma
| pong
| tyingq wrote:
| http://anirudhjoshi.github.io/fluid_table_tennis/
|
| No good for mobile, but perhaps someone could fix that.
| exikyut wrote:
| There's a paid Android app linked from that page. You are
| technically correct that the [free] webpage indeed not work
| on mobile.
| LoSboccacc wrote:
| I'm fine with non free thanks both for the links and the
| tips!
| faitswulff wrote:
| I truly miss this game sometimes. Is it still available
| anywhere?
| dspillett wrote:
| The rights owners for pong (Atari?) had it and any mirrors
| they could find taken down.
|
| A quick search suggests that there are downloadable copies
| out there. 1.3c was the latest release. Check WayBackMachine
| from 2007-Sep-13 (https://web.archive.org/web/20070913204145/
| http://www.plasma...) or older if you want more info, after
| that the site is an "Atari told us to take this down" message
| for a while and is currently held by someone else talking
| about other games.
___________________________________________________________________
(page generated 2020-08-05 23:01 UTC)