InAI AdvancesbySascha Kirch·Apr 6Making SAM3 8x Faster — What the Profiler Actually ShowedWhat profiling a state-of-the-art segmentation model teaches you about running research code efficientlyA response icon2A response icon2
InAI AdvancesbySascha Kirch·Mar 1Inside the CUDA Kernel: The GPU Implementation of 3D Gaussian SplattingBonus Part: How Flat Arrays, Radix Sort, and Shared Memory Turn Conceptual Tiling Into 100+ FPS RenderingA response icon1A response icon1
InAI AdvancesbySascha Kirch·Feb 14The Tricks That Make Production 3DGS Fast (Even If Ours Isn’t)Part 4 of Building a 3D Gaussian Splatting Renderer — Understanding the Optimizations That Make Real-Time Rendering PossibleA response icon2A response icon2
InAI AdvancesbySascha Kirch·Feb 7Splat Your Own Gaussians: From Circles to EllipsesPart 3 of Building a 3D Gaussian Splatting Renderer —Covariance Matrices, Jacobians, and Proper 2D ProjectionA response icon2A response icon2
InAI AdvancesbySascha Kirch·Feb 1Circles Are Not Gaussians (But Let’s Pretend They Are)Part 2 of Building a 3D Gaussian Splatting Renderer — Circles, Colors, and TransparencyA response icon2A response icon2
InAI AdvancesbySascha Kirch·Jan 25I Built the Slowest 3D Gaussian Splatting Renderer… On PurposeBuilding a deliberately slow, crystal-clear Python renderer to reveal the implementation details that papers skip and GPU kernels hideA response icon5A response icon5
InAI AdvancesbySascha Kirch·May 19, 2025Still Avoiding einsum()? It’s Time to Fix ThatBuild Intuition, Master the Syntax, and Apply einsum() with Real Examples and a Handy Cheat SheetA response icon6A response icon6
InAI AdvancesbySascha Kirch·May 12, 2025Mastering NumPy: Manual Metadata Manipulation for Memory-Efficient ArraysPart 3: NumPy Like a Pro: A Deep Dive Into Arrays and PerformanceA response icon2A response icon2
InAI AdvancesbySascha Kirch·May 4, 2025The Power of Views: How NumPy Avoids Copies and Saves MemoryPart 2: NumPy Like a Pro: A Deep Dive Into Arrays and PerformanceA response icon5A response icon5
InAI AdvancesbySascha Kirch·Apr 28, 2025Why NumPy Arrays Are So Fast (And How They Really Work)Part 1: NumPy Like a Pro: A Deep Dive Into Arrays and PerformanceA response icon5A response icon5