|
That's delightfully (deliciously!) satisfying and soothing to watch.
|
# ? Feb 26, 2017 02:11 |
|
|
# ? Jun 8, 2024 06:09 |
|
pokeyman posted:That looks awesome! Anything interesting to share about the implementation? (I know absolutely nothing about fluid simulation.) Tres Burritos posted:Man the only thing giving that away is the weird jagged edges. Thanks! I noticed the jaggyness after simulating the animation. The reason is because the fluid is being dropped on a surface that is tilted slightly towards the camera. The simulator works by making calculations on a 3D grid, and because of this, completely smooth slopes aren't able to be represented with 100% accuracy. It's kind of like the fluid is falling down tiny little stair steps, which is what is causing the choppiness. The choppiness could probably be reduced by tweaking a few settings. Or I could have just dropped the fluid on a flat surface that is aligned to the simulation grid. Here is a visualization of how the simulator 'sees' the sloped surface. Notice the 'stairstep' banding artifacts.
|
# ? Feb 26, 2017 02:17 |
|
RexJericho posted:Here is a visualization of how the simulator 'sees' the sloped surface. Notice the 'stairstep' banding artifacts. Very cool. How long does it take to run the simulation?
|
# ? Feb 26, 2017 11:07 |
|
munce posted:Very cool. How long does it take to run the simulation? This simulation/animation took a total of 49.4 hours for 300 frames. 1.5 hours to simulate, 7.4 hours to convert the particles into triangle meshes, and 40.5 hours to render into an animation. My computer isn't very powerful and is quite slow for simulating/rendering. I used an ultrabook style laptop with Intel Core i5-4200U @ 1.60GHz processor, integrated Intel HD4400 graphics chip, and 8GB RAM.
|
# ? Feb 26, 2017 20:49 |
|
RexJericho posted:This simulation/animation took a total of 49.4 hours for 300 frames. 1.5 hours to simulate, 7.4 hours to convert the particles into triangle meshes, and 40.5 hours to render into an animation.
|
# ? Feb 27, 2017 13:56 |
|
RexJericho posted:This simulation/animation took a total of 49.4 hours for 300 frames. 1.5 hours to simulate, 7.4 hours to convert the particles into triangle meshes, and 40.5 hours to render into an animation. Oh god, no wonder it took so long to render.
|
# ? Feb 27, 2017 14:01 |
|
Tres Burritos posted:Man the only thing giving that away is the weird jagged edges. I would have put that down to bad compression on the video
|
# ? Feb 27, 2017 14:18 |
|
That is the best possible response. Godspeed RexJericho.
|
# ? Feb 27, 2017 14:22 |
|
MrMoo posted:Dynamic CSS and Polymer is surprisingly challenging. Chrome pulled CSS variables in v49 and now they have reintroduced a subset via a Polyfill, I was hoping to use CSS like color: var(--tick-colour); but due to performance reasons the variable is never recalculated automatically. Why would you need anything dynamic for color? Don't they just scroll off into oblivion and come back with different numbers later?
|
# ? Feb 27, 2017 15:55 |
|
Worked on a little LED messenger bag display thing probably more useful for cyclists to piss off drivers but I enjoy it: https://twitter.com/lindseybieda/status/834225556816326659 https://twitter.com/lindseybieda/status/835192084395077634 Code and how to on github.
|
# ? Feb 28, 2017 00:52 |
|
That's fun, you should try to get voice control in there for when you're on the go
|
# ? Feb 28, 2017 02:03 |
|
Munkeymon posted:Why would you need anything dynamic for color? Don't they just scroll off into oblivion and come back with different numbers later? This isn't the ticker, this is famous Twitter screens, prices go up and for them generally down and thus need to change from green to red.
|
# ? Feb 28, 2017 04:19 |
|
clockwork automaton posted:Worked on a little LED messenger bag display thing probably more useful for cyclists to piss off drivers but I enjoy it: And you got it running on Chrome too
|
# ? Feb 28, 2017 13:23 |
|
I'm sick so I've been killing time by messing with the 286 card in my A2000. I installed a VT102 emulator on it and rigged up a C# program to get WMI data from OpenHardwareMonitor and spit it out over serial with VT100 escape sequences. When the temps get too high they blink red!
|
# ? Mar 1, 2017 04:47 |
|
Waiting for IPOs is a bit tedious ... $21/23 pricing it looks like. MrMoo fucked around with this message at 16:06 on Mar 2, 2017 |
# ? Mar 2, 2017 16:01 |
|
e: nevermind, not ready to share.
Knyteguy fucked around with this message at 18:20 on Mar 2, 2017 |
# ? Mar 2, 2017 18:13 |
|
For about the past month, following some conversations with a friend who works in the sciences, I've been working on a toolkit. Its purpose is to let people with only basic knowledge of Linux, and no specific knowledge of distributed filesystems, create a durable in-house data archival solution out of spare PCs. This photo is of the cluster converging after the first successful bring-up with no manual intervention (so, basically, a trouble-free guided install). There's a lot of work to do to make this more than a glorified installer, but I hope it helps some people.
|
# ? Mar 8, 2017 04:13 |
|
MrMoo posted:My nephew barely knows any HTML or CSS but has managed to make some progress copying the NYSE floor designs, I updated with Polymer and WebComponents to hook up with WebSocket streamed live data. That looks really nice, but just a fair warning that Polymer is a flaming pile of garbage. So many of the bundled elements are broken in extraordinary ways, especially when you combine them and try to make it compatible for different browsers. It ends up being way more trouble than it is worth.
|
# ? Mar 8, 2017 18:00 |
|
I blew the dust off of an old 6502 disassembler I wrote and spent two evenings turning it into a NES-worthy cycle-accurate emulation core: I've added just enough PPU functionality in there to set the PPU status register flags and trigger NMIs to keep the execution from looping forever on vblanks and the like for Super Mario Bros.
|
# ? Mar 8, 2017 18:45 |
|
hendersa posted:I blew the dust off of an old 6502 disassembler I wrote and spent two evenings turning it into a NES-worthy cycle-accurate emulation core: I got as far as setting up the PPU flags before losing interest in my emulator.
|
# ? Mar 8, 2017 18:57 |
|
LP0 ON FIRE posted:That looks really nice, but just a fair warning that Polymer is a flaming pile of garbage. So many of the bundled elements are broken in extraordinary ways, especially when you combine them and try to make it compatible for different browsers. It ends up being way more trouble than it is worth. yes Polymer is garbage. I've spent years working on an app that was all Polymer and managed billions of dollars of assets. It was an unworkable mess getting Web Components to play nice. e: read more about the project MrMoo posted:Dynamic CSS and Polymer is surprisingly challenging. Chrome pulled CSS variables in v49 and now they have reintroduced a subset via a Polyfill, I was hoping to use CSS like color: var(--tick-colour); but due to performance reasons the variable is never recalculated automatically. I built something extremely similar to this at an old job, it wasn't stock prices but a different financial instrument monitor, complete with dynamic colors/styles based on data flow changes. Use Angular Directives instead of Polymer elements. It will play much nicer with data binding and dynamic styling. Careful Drums fucked around with this message at 19:04 on Mar 8, 2017 |
# ? Mar 8, 2017 18:57 |
|
mdxi posted:
That's a super cool idea and I would love for it to work. I have tons of computers laying around and nowhere (other than the cloud) to store family photos and poo poo.
|
# ? Mar 8, 2017 19:00 |
|
Careful Drums posted:That's a super cool idea and I would love for it to work. I have tons of computers laying around and nowhere (other than the cloud) to store family photos and poo poo. Just remember that on-site replication isn't 'backup'
|
# ? Mar 8, 2017 19:31 |
|
Munkeymon posted:Just remember that on-site replication isn't 'backup' Yeah. Ideally I'd like to store all my stuff locally with 'cold' online storage.
|
# ? Mar 8, 2017 19:50 |
|
Careful Drums posted:yes Polymer is garbage. Christ that sounds stressful. I've been working on something that's being tested at 30 different schools that uses Polymer. When it's official, I think rebuilding it with Angular will be the way to go.
|
# ? Mar 8, 2017 20:17 |
|
Luigi Thirty posted:
I can see why. The PPU guts are pretty awful. Honestly, there is no shame in bailing out at that point. The 6502 core is arguably the most reusable bit of it all, anyway.
|
# ? Mar 8, 2017 21:24 |
|
hendersa posted:I can see why. The PPU guts are pretty awful. Honestly, there is no shame in bailing out at that point. The 6502 core is arguably the most reusable bit of it all, anyway. Yeah, it started as an Apple I emulator and I tried bolting a PPU onto it instead of the video terminal. That was easy at least since the Apple spits the character it's outputting to a memory-mapped I/O register and doesn't do anything more than teletype output. I could load BASIC from ROM and use it to load cassette dumps. Luigi Thirty fucked around with this message at 00:07 on Mar 9, 2017 |
# ? Mar 9, 2017 00:05 |
|
Careful Drums posted:yes Polymer is garbage. Thomson Reuters EIkon is planning to use Polymer everywhere, from a diaspora of every JS framework in existence, including home grown, and from a distance looks really convenient and neat, they've been building up quite an extensive web components library for a while. Due to CEF they lock into one browser and myself at NYSE all the end displays are a custom build of Chromium so that's thankfully not much of a concern. I only have a handful of screen types to work with and pretty much done already so I've been burnt already on all its failings. I think the performance angle from declarative polymers is going to be quite useful and is reasonably well defined and understandable view in the system. I need to spend a bit more time on the infrastructure, currently juggling between Node, TypeScript, C++ × Beast, Java × JSR 356, Elixir × Cowboy, Python3 × Async, and now looking at Erlang. So the plethora of frontend frameworks has to sit on the TODO list, I get to meet a UX designer on Friday so see if she has any preference.
|
# ? Mar 9, 2017 00:22 |
|
Hey, I've been working on a weird dealio for predicting how long github issues will take (like a super small version of COCOMO, but not a giant piece of poo poo and with no human input). Weirdly complicated as a machine learning dealio. It's not incredible but it's not bad either, and I've been wondering if it was better than humans. Would any of you be up for a go? (completely command line as of now so nothing to show really)
|
# ? Mar 9, 2017 06:51 |
|
Careful Drums posted:That's a super cool idea and I would love for it to work. I have tons of computers laying around and nowhere (other than the cloud) to store family photos and poo poo. It actually does work, as of right now. There's just no tooling yet to make things like swapping dead drives or reimaging OS partitions easy -- so at this moment you get a storage cluster built for you in an automated way, but then maintaining it requires lots of domain knowledge. Mechanizing common repairs is one of my next major tasks. Edit: Also, the partitioning layout will be improved in the next milestone, and cluster initialization will use a more efficient (but equally durable) storage algorithm, as long as you have at least 5 machines in the cluster. So while it works just fine, pretty much everyone should wait for the next release. Munkeymon posted:Just remember that on-site replication isn't 'backup' Absolutely, but some people I want to help have data that they can't (or shouldn't, or don't want to) put on quasi-public utilities like S3 or Glacier. Or possibly can't afford to do so (science types get paid very, very little in many cases). If you're a technologist with discretionary income and want to set up an in-house storage cluster, I completely agree that you should be periodically mirroring it to Glacier or similar. mdxi fucked around with this message at 07:35 on Mar 9, 2017 |
# ? Mar 9, 2017 07:30 |
|
curufinor posted:Hey, I've been working on a weird dealio for predicting how long github issues will take (like a super small version of COCOMO, but not a giant piece of poo poo and with no human input). Weirdly complicated as a machine learning dealio. It's not incredible but it's not bad either, and I've been wondering if it was better than humans. Would any of you be up for a go? This might not be the high-traffic thread to ask in! Maybe the general thread, or the agile one or something?
|
# ? Mar 9, 2017 16:46 |
|
mdxi posted:Absolutely, but some people I want to help have data that they can't (or shouldn't, or don't want to) put on quasi-public utilities like S3 or Glacier. Or possibly can't afford to do so (science types get paid very, very little in many cases). Yeah, and I didn't mean to impugn your project or anything, either - I see why it'd be useful to anyone without much budget for new hardware. Heck, I might use it eventually if I ever get around to both fixing my NAS and building a second one to replicate to like I want to
|
# ? Mar 9, 2017 16:47 |
I'm fiddling with a prototype platformer that borrows everything I like about Gunbound, Fez, Shenzen I/O, and Super Metroid. Most of the devices have simple CPUs which you can program in an assembly language. I'm having fun with it so far. It's a crazy niche market, though, so we'll see how far it goes.
|
|
# ? Mar 10, 2017 17:51 |
|
Implementing features and not caring about accurate styles produces some interesting results. I've actually got pretty far with Elixir too, but I'm seeing only IPv4 with Cowboy and error messages still blow. Polymer 2 is almost out too.
|
# ? Mar 13, 2017 04:43 |
|
I updated my bag so I can now charge it while powering it which is definitely helpful when I need it on for longer periods of time. I'm trying to figure out an issue with the bluetooth sending now where sometimes the first letter of a message I send via bluetooth is missing, but only the first letter.
|
# ? Mar 14, 2017 03:25 |
I added buffers which allow me to connect CPUs to each other. They're simple objects which inherit from Line2D and directly copy bytes of memory from one device to another. Means I can memory map devices. I'm also training a chatbot in the background so, ideally, each segment will have three paths the player can take: the dialog path where they need to negotiate with the AI, the programming path where they need to solve a simple puzzle, and the exploration path where they need to look around.
|
|
# ? Mar 30, 2017 06:43 |
|
Jo posted:
My all time favorite thing to code (or theorize and then never get around to doing,) is wiring and cpu simulations. I gotta do something.
|
# ? Mar 30, 2017 22:07 |
|
Now with added Gotham, looks pretty.
|
# ? Apr 2, 2017 14:49 |
|
Updated with remote admin capability so the contents can be updated with all the different screen types:
|
# ? Apr 6, 2017 03:48 |
|
|
# ? Jun 8, 2024 06:09 |
|
Once again I failed to get a job that I applied for and I ended up depressed, meaning I found a dumb project to occupy myself. I found a detailed description of the Pac-Man arcade hardware and a Z80 assembler... Not bad for half an hour of work
|
# ? Apr 8, 2017 06:15 |