iiinova — where your digital self keeps evolving

Geolm @geolm.bsky.social · 09/10/2026

my new lib bc_packed is out! currently BC1 only, 1.47:1 average compression ratio, CPU/GPU parallel decompression (15300 MiB/s on my M5 Pro GPU), rice encoding with static rank tables. lossless and no dependencies! github.com/Geolm/bc_pac...

github.com

GitHub - Geolm/bc_packed: An asymmetric lossless compressor for already-compressed BC textures, designed for full GPU decompression.

An asymmetric lossless compressor for already-compressed BC textures, designed for full GPU decompression. - Geolm/bc_packed

020

Geolm @geolm.bsky.social · 07/10/2026

12654.0 MiB/s on my m2 max for 1000 bc1 1024x1024 texture decompression... mainly bottleneck by the bandwidth. Anyway nice boost :)

010

Geolm @geolm.bsky.social · 06/10/2026

I'm about to archive my bc_crunch library as I will release soon its successor that focuses on high-speed decompression. Latest test gives 2650.8 MiB/s on CPU but the decoder is designed to run on the GPU. I hope I can share some GPU numbers soon...

120

Geolm @geolm.bsky.social · 25/09/2026

it's weird but somehow my native webgpu version of onedraw (github.com/Geolm/onedra...) is less popular than the Metal's one. Are you all on macOS? ;) By the way, I'm still adding some features (I'm guilty of having too many projects at the same time).

github.com

GitHub - Geolm/onedraw-webgpu: WebGPU tile-based 2d renderer with compute-driven binning and indirect draws

WebGPU tile-based 2d renderer with compute-driven binning and indirect draws - Geolm/onedraw-webgpu

010

Geolm @geolm.bsky.social · 14/06/2026

Forgot to mention: my GPU-driven 2D WebGPU renderer is now public: github.com/Geolm/onedra... Not an official release yet, but it's usable. One .c file to integrate, tested on Win32 and macOS. More features coming once the heat wave stops melting me.

github.com

GitHub - Geolm/onedraw-webgpu: WebGPU tile-based 2d renderer with compute-driven binning and indirect draws

WebGPU tile-based 2d renderer with compute-driven binning and indirect draws - Geolm/onedraw-webgpu

030

Geolm @geolm.bsky.social · 20/05/2026

it really depends on the project IMO I write mainly rendering, compression, low-level stuff so in my case I can work around the limitations and publish code that can be used by a lot of languages since it's in C

010

Geolm @geolm.bsky.social · 19/05/2026

I used to miss "generics" in C but now I only use arrays in C and I'm fine :)

100

Geolm @geolm.bsky.social · 05/05/2026

it won't be supported on macOS I suppose....

100

Reposted by Geolm

Markus Schütz @m-schuetz.bsky.social · 24/04/2026

New Paper🙂 Nanite has shown that small triangles can be rendered fast in compute, we're exploring how fast for large meshes with up to 18.9 billion triangles, without the need to precompute LOD structures. Paper: github.com/m-schuetz/Cu... Source: github.com/m-schuetz/Cu...

26820

Geolm @geolm.bsky.social · 31/03/2026

Some news about my WebGPU gpu-driven 2d port: The GPU binning phase is working, but there are still issues with culling, likely related to AABB tests. Debugging WebGPU in Xcode is a nightmare: a lot of passes unneeded, no way to inspect shader source or code easily, no stepping in shaders...

000

Geolm @geolm.bsky.social · 20/03/2026

I’m porting my 2D SDF GPU-driven renderer (onedraw) from Metal to the WebGPU API. A few observations so far: - Pure C API, no more Objective-C or metal-cpp 👍 - WGSL :( - A lot more boilerplate for bind group and pipeline setup - ~30 MB executable just to clear the backbuffer

011

Geolm @geolm.bsky.social · 23/01/2026

As people asked for better decompression speed in my bc_crunch lib, I started experimenting with Huffman coding. First results look promising: compression ratio is still better than zlib (~1.52:1) and decompression reaches ~144 MB/s. github.com/Geolm/bc_cru... (BC1 only for now)

github.com

GitHub - Geolm/bc_crunch at huffman

tiny dependency-free lossless compressor for BC/DXT texture streams (C99, ~1000 LoC) - GitHub - Geolm/bc_crunch at huffman

040

Geolm @geolm.bsky.social · 13/01/2026

nope, as you said it's complicated to get it done properly, I don't have the need ATM (not fan of the property grid everything) and it would add a lot of LOC to me lean lib.

000

Geolm @geolm.bsky.social · 07/01/2026

I didn't had the luxury to work on my personal projects during the holidays season but here's a video of my attempt to animate my creature using position based dynamics (distance/angle/elbow constraint + ik).

010

Geolm @geolm.bsky.social · 14/12/2025

also with multiple quads and alpha blending you have more overdraw and r/w access to the framebuffer compared to tile rendering. Another point is primitive smoothin and boolean operation that are only possible with tile

010

Geolm @geolm.bsky.social · 12/12/2025

got a 404 error when trying to access the pdf...

100

Geolm @geolm.bsky.social · 12/12/2025

People kept asking me to compare bc_crunch with zlib, so I finally ran the tests. Overall, bc_crunch comes out ahead. More details are available on the GitHub page. bc_crunch BC1: 1.580368 vs zlib: 1.436144 bc_crunch BC4: 1.327702 vs zlib: 1.104125 I used miniz with the highest compression setting

010

Geolm @geolm.bsky.social · 10/12/2025

another note : BC5 ratio is pretty bad at the moment because it treats the image as two independent channels. One solution would be to create a new function only for normalmaps that exploit inter-channel prediction (not planned for now).

000

Geolm @geolm.bsky.social · 10/12/2025

btw I also found some small improvements for BC4, ratio went from 1.23 to 1.29

100

Geolm @geolm.bsky.social · 10/12/2025

I understand, I might need that at some point too, if so I'll update the lib.

010

Geolm @geolm.bsky.social · 10/12/2025

I don't know, I did not try nanovg. Onedraw supports less features for sure. I wrote this lib for my own needs and to try some ideas I had in mind, not to compete with others libs.

110

Geolm @geolm.bsky.social · 10/12/2025

Hey, I thought I was done with bc_crunch, but I tried one last idea: Inter-channel differential coding. It works pretty well. The average compression ratio for BC1 went from 1.494:1 to 1.580:1, finally breaking the 1.5× barrier. It's already available here: github.com/Geolm/bc_cru...

github.com

GitHub - Geolm/bc_crunch: tiny dependency-free lossless compressor for BC/DXT texture streams (C99, ~800 LoC)

tiny dependency-free lossless compressor for BC/DXT texture streams (C99, ~800 LoC) - Geolm/bc_crunch

140

Geolm @geolm.bsky.social · 08/12/2025

Do you mean using a texture map as a probability field for spawning? I'm curious, what would be the use case?

100

Reposted by Geolm

Walaber @walaber.com · 04/12/2025

Haven't done a camera analysis in a while! Here's a quick look at the camera for Metroid Dread 🎥

5437102

Geolm @geolm.bsky.social · 03/12/2025

Experimenting with a new control: a knob :) It should land in the public leanUI library soon. And yes, I’m working on a synth… stay tuned (pun intended).

100

Geolm @geolm.bsky.social · 03/12/2025

As a Canadian, I can tell you the shovel sound is a bit off. You only get that bright metal-on-concrete noise when the ground is bare. Most of the time there’s a layer of packed snow or ice stuck to the concrete, which makes the shovel sound more muted and less metallic.

220

Geolm @geolm.bsky.social · 30/11/2025

no the post can't be that long !! ;)

020

Geolm @geolm.bsky.social · 27/11/2025

Needed a simple way to spawn 2D objects (grass blades, stones, etc.) in composite shapes or along splines, so I wrote this: github.com/Geolm/spawn2d. It’s minimal, single-file, and easy to drop into any project.

github.com

GitHub - Geolm/spawn2d: Uniform sampling in combined shapes with overlap handling and distance constraints. (~300 LOC, C11)

Uniform sampling in combined shapes with overlap handling and distance constraints. (~300 LOC, C11) - Geolm/spawn2d

131

Geolm @geolm.bsky.social · 26/11/2025

Uniform Catmull–Rom spline sampling with disc collision and exclusion zones. Poisson-like, but curve-specific.

020

Geolm @geolm.bsky.social · 26/11/2025

Working on a uniform sampler with dynamic exclusion zones for procedural gen. Solid results so far, code release is possible once it’s cleaner.

010

Geolm @geolm.bsky.social · 21/11/2025

currently no, TBH it's not planned since my goal was to have a quick way to export video from my games/renderer. Feel free to fork the repo and add the features you need.

010

Geolm @geolm.bsky.social · 21/11/2025

Just released a single-file C99 library to create MJPEG AVI videos from an RGBA framebuffer. github.com/Geolm/mjpegw Like jo_mpeg, but MJPEG gives higher quality (larger files). • 3-function API • No dependencies • No audio • Plays in ffmpeg & VLC Huge thanks to Tiny-JPEG for the JPEG part.

github.com

GitHub - Geolm/mjpegw: Tiny C99 library for creating MJPEG AVI files. No external dependencies

Tiny C99 library for creating MJPEG AVI files. No external dependencies - Geolm/mjpegw

130

Geolm @geolm.bsky.social · 19/11/2025

Also, there’s a benchmark in the repo at /test/benchmark.c that profiles BC1 crunch/decrunch—make sure to compile with optimization flags.

000

Geolm @geolm.bsky.social · 19/11/2025

Some optimizations news: NEON + decoder table in bc_crunch → decrunch ~1.5× faster, crunch ~1.3× faster. No plans for more performance tweaks right now, but I left some ideas in the README. Ping me if you want to chat about them!

100

Geolm @geolm.bsky.social · 18/11/2025

Hi everyone! Thanks for the interest in my library bc_crunch. Quick update: I’m currently working on optimizations. Two major bottlenecks have been identified, and fixes are on the way. I’ll share some performance numbers soon and add a profiling test to the repo.

020

Geolm @geolm.bsky.social · 17/11/2025

I’ve updated the BC4 algorithm and I’m curious whether it improves the compression ratio on your textures as well. For reference, my AO/displacement map went from 1.051032 to 1.211644 (a 15.3% gain). I have still some ideas left, I will try to continue to improve more.

000

Geolm @geolm.bsky.social · 17/11/2025

Update, I changed the heuristic for BC4 streams, went from an average ratio of 1.051032 to 1.211644. Tested on AO, displacement map (check textures/bc4 folder)

000

Geolm @geolm.bsky.social · 17/11/2025

Thanks! I haven't had time to profile decompression yet, so that measurement is really helpful. Would you mind if I add that number to the project page?

100

Geolm @geolm.bsky.social · 17/11/2025

Thanks for the feedback! I agree on BC4/5 — right now it’s a bit hit-or-miss and too dependent on input noise. The algorithm still needs improvement there. I’m currently building a BC5 test texture set, which should help compare different approaches more reliably.

000

Geolm @geolm.bsky.social · 15/11/2025

Don't hesitate to message me if you find a bug or if you have a question or feedback. I will update the readme with the BC4 benchmark soon.

010

Geolm @geolm.bsky.social · 15/11/2025

Just released bc_crunch: a tiny (~700 LOC), C99 library for lossless BC1/BC4/BC3/BC5 compression. One .c/.h pair, no deps, no malloc GPU-ready output Zigzag + delta + popcount/dictionary tricks Tested Perfect for games, procedural assets, streaming textures. github.com/Geolm/bc_cru...

github.com

GitHub - Geolm/bc_crunch: tiny dependency-free lossless compressor for BC/DXT texture streams (C99, ~700 LoC)

tiny dependency-free lossless compressor for BC/DXT texture streams (C99, ~700 LoC) - Geolm/bc_crunch

3408

Geolm @geolm.bsky.social · 12/11/2025

Hey everyone! I’m working on bc_crunch, a lossless BC1 (DXT) texture compressor/decompressor in C99 (~500 LOC, no dependencies). It compresses and decompresses textures bit-for-bit. Thinking of sharing it publicly, along with a doc on techniques tried—anyone interested?

000

Geolm @geolm.bsky.social · 11/11/2025

thank you for the feedback :)

010

Geolm @geolm.bsky.social · 07/11/2025

Big thanks to @flohofwoe.bsky.social for sokol_app — made it easy to keep github.com/Geolm/onedraw examples small and clean, no extra dependencies. Super smooth setup. Only thing I couldn’t find was a way to create sRGB backbuffer on macOS (did I miss that in the docs?).

github.com

GitHub - Geolm/onedraw: GPU driven 2d sdf renderer drop-in library

GPU driven 2d sdf renderer drop-in library. Contribute to Geolm/onedraw development by creating an account on GitHub.

171

Geolm @geolm.bsky.social · 04/11/2025

Excuse the coder art — this GIF shows the tile culling debug mode. Blue tiles are the ones that make it to the rasterizer. Notice how the interior of hollow shapes is efficiently skipped. You can have look at the code and documentation here: github.com/Geolm/onedraw

000

Geolm @geolm.bsky.social · 03/11/2025

Yes brute may work, but actually I'm trying to achieve good performances and use the "right" amount of sdf shapes. My test involved a bit, now I'm using capsule (blue)+box (green, for the sharp endpoints) and biarc (red). Still to handle the worst case (all control points colinear though)

010

Geolm @geolm.bsky.social · 03/11/2025

bsky only used the first image of my apng so let's use gif like it's 1998

010

Geolm @geolm.bsky.social · 03/11/2025

Testing an approach to break a quadratic bezier into small SDFs: Blue: straight lines (oriented box SDF) Green: “pie” caps connecting lines Red: arcs Faster than a full bezier SDF, though tricky with large stroke widths. All done with my open-source GPU renderer github.com/Geolm/onedraw

240

Geolm @geolm.bsky.social · 03/11/2025

Hello everyone, I wrote an extensive documentation on the GPU‑driven 2D renderer I’ve been working on. Check out the overview here: github.com/Geolm/onedra... If you're interested, here's the repo URL: github.com/Geolm/onedraw #GPU #2DGraphics #OpenSource #Metal #Rendering

160

Geolm @geolm.bsky.social · 01/11/2025

Calling macOS (Apple Silicon) devs Looking for early testers for onedraw, my GPU-driven 2D renderer built with Metal. Feedback before release would be super appreciated 🙏 github.com/Geolm/onedraw #Metal #GPUDriven

github.com

000