- presents an efficient appearance framework and a compact neural model for 3D Gaussian Splatting
- compares spherical harmonics and recent spherical appearance models
- shows how the framework allows reducing primitive cost and improving reconstruction quality
- releases version 2 of Alwan, a color science library written in C11
- provides color space conversions, chromatic adaptation, transfer functions, ACES pipelines, color appearance models, gamut mapping, and spectral tools
- offers dual f32/f64 precision, SIMD-accelerated bulk kernels, and an optional byte-exact deterministic mode
- provides validation against reference implementations
- presents a self-contained C++ experiment that encodes RGB images or multi-layer PBR materials as a shared compressed texture plus a tiny MLP decoder
- trains the latent texture and the decoder entirely with Evolution Strategies
- provides a learning roadmap covering the three pillars of machine learning mathematics: linear algebra, calculus, and probability theory
- walks through vectors and vector spaces, norms, bases, linear transformations, matrices, determinants, eigenvalues, and matrix decompositions
- continues with single- and multivariable calculus, gradients, etc., and probability topics
- implements the ideas of the “No Graphics API” SIGGRAPH talk as a thin Vulkan 1.4 wrapper
- eliminates public buffer objects, descriptor sets, pipeline layouts, and sampler objects from the API
- includes Windows examples
- additionally documents hardware support across AMD and NVIDIA GPUs
- documents the first three steps of building an optimized BF16 GEMM kernel in CuTeDSL
- walks from a naive 4 TFLOP/s to finally reaching 285 TFLOP/s
- explains CuTeDSL tensors, layouts, tiles, atoms, and the Blackwell tcgen05 MMA and tensor memory architecture
- explains how to write custom post-processing effects for Unity’s Universal Render Pipeline using Render Graph
- implements a greyscale effect, a depth-based silhouette effect, and an outline effect
- interview with Neil Trevett covering Khronos standards including Vulkan, SPIR-V, Slang, OpenXR, glTF, and ANARI
- Discusses Vulkan’s push toward usability with descriptor heaps, cooperative matrices/vectors for ML inference, and Linux gaming momentum
- also covers Slang’s adoption, OpenXR’s expansion into smart glasses, the Open Metaverse Browser initiative, and others
- explains volume rendering as a set of techniques for visualizing 3D scalar data stored in a voxel grid
- shows applications in various fields, such as medical imaging and visual effects (e.g., clouds, smoke, fire, and fog)
- demonstrates examples including 3D texture slicing and colormap classification
Thanks to Graham Wihlidal for support of this series.
Would you like to see your name here too? Become a Patreon of this series.
