Guilherme Avila retweeted
New blog post! "Load store conflicts", in which we look at some performance sensitive code that has surprisingly dramatic performance swings based on the compiler and the microarchitecture used. RTs appreciated!
zeux.io/2025/05/03/load-stor…
Guilherme Avila retweeted
GH stars / retweets welcome as usual ;)
GitHub release link: github.com/zeux/meshoptimize…
Very cool Garrett, transition looks on point.
I find ortho projection for 3D map visualizations a bit jarring, but many professionals in carthography & construction seem to prefer it. Having a smooth transition between projections, helps a lot keeping track of what's going on.
Makes sense, didn't think about the need to adjust position so the frustum aligns. And, yes. I don't see any jitter at well. Well done!
very interesting looking effect indeed, are you animating the perspective FOV down to zero and transitioning to ortho? I thought about something similar, but I was worried it would cause visual artifacts, cause low FOVs tend to be a bit finnicky.
follow-up optimization
xitter.cf/ScieCode/status/181108…
Found another potential optimization, breaking up instruction dependency, and allowing better pipeline while quarter-rate rcp runs.
Got mixed results here, either no net gain or ~5% gain. Probably depends on the compiler, but should use this formulation eitherway.
It has been a net gain in most hardwares I've managed to get my hands on, but I only have access to a small set of gpus. Please let me know how your test goes.
Worth noting, snippet expects unit normals & returns uv in [-1,1] range (remap to [0,1] afterwards if the API expects that range)
I see Stubbe's octahedral decode opt in almost every usage I come across. But I've never seen it used for encoding.
The same symmetry applies here! and its ~50% faster.
Especially important for octahedral-encoded textures, as it's required for every sample! please RT.
Guilherme Avila retweeted
meshoptimizer v0.21 is out!
This release has many improvements related to meshlets: a new function to optimize meshlets for HW efficiency, extra features for meshopt_simplify* to make Nanite style processing easier to implement and faster, and more! RTs and ⭐️ appreciated :)
Octahedral-encoded indirect specular map, with added border so it allows bilinear filtering, pre-convolved mips & trivially tiled within texture atlas.
shadertoy.com/view/43G3Dd
