Save Detail Attributes in Houdini .bgeo Cache Files

A full SPICE circuit simulator. In your browser. No install. 👉 https://lnkd.in/geyZp2pp Drop in a KiCad .kicad_sch file, click components to set SPICE models, place probes, run. That's it. The ngspice engine is compiled to WebAssembly and runs locally — first load grabs ~20 MB, then everything is client-side. Built with React + TypeScript + eecircuit-engine. Why: I wanted something modern where I could paste .model text straight from a manufacturer datasheet — or from an LLM. I've been using Claude to generate component models since I don't know SPICE . Component settings bind to RefDes and save to JSON, so re-importing an updated schematic keeps your simulation setup intact.

For people doing computational materials / atomistic simulation work with VASP, I wanted to share a small side project that grew out of my own workflow frustrations while working with structure files inside VS Code. I made a lightweight VS Code extension called AtomView for viewing POSCAR, CONTCAR, and XDATCAR files directly inside the editor. The idea is not to replace tools like VESTA, but to make quick inspection and sanity-checking a bit smoother during day-to-day workflows without constantly switching between editing structures and opening external viewers. Current features include: - live auto-updating structure preview - animated trajectory playback for XDATCAR - atom hover coordinate inspection - selective dynamics highlighting - bond rendering Hopefully this might be useful to someone else with a similar workflow! https://lnkd.in/gCxkDaz2

Quick rendering math lesson (common trap): FoV in Unreal Engine isn't interpolated correctly when blending between two cameras The current code does: FoV = lerp(FoVA, FoVB, Weight) however, that's not how you interpolate FoV! To be "perceptually" linear and also correct under aspect ratio scaling, FoV needs to be lerped in tangent space: FoV = atan(lerp(tan(FoVA*0.5), tan(FoVB*0.5), Weight)) * 2 The relationship between FoV and aspect ratio isn’t linear, because FoV is angular! Video showcasing the problem: https://lnkd.in/egAs6QYB Blending between a 32:9 camera with 98.21° FoV and a 16:9 camera with 60° FoV. Viewport is 32:9. Due to FoV scaling being Hor+ by default in UE, at 32:9 they should look identical. But during blends, they don't (the FoV isn't stable)! In practice, it also means camera interps aren't as smooth as they could be; the attached video shows that. PR to fix the problem: https://lnkd.in/eXtQNNiT hopefully it gets merged!

What happens when a field meets a boundary? I built KNOT Boundary Field Lab: a browser-based simulation lab for waves, nonlinear fields, reaction-diffusion, pattern tracking, replay, and testing. It is not magic. It is not claiming particles. It is a tool for observing emergence under constraint. https://lnkd.in/d6EDnbbP

#Day60 of DSA Journey | LeetCode Challenge https://lnkd.in/du5CBBXH Problem: Rotating the Box (Medium) Day 60! Today’s problem was a great combo of matrix transformation + simulation (gravity effect). Understanding the Problem: We are given a box with: # → stones * → obstacles . → empty spaces The box is rotated 90° clockwise, and then: Stones fall down due to gravity Obstacles stay fixed Key Insight: Instead of directly rotating + simulating together, break it into 2 steps: Simulate gravity row-wise (before rotation) Move stones # to the right (like falling down after rotation) Rotate the matrix (transpose + reverse rows) Approach: For each row: Start from right

At the beginning, what we now call the Finite Element Method (FEM) started as something much simpler: the Direct Stiffness Method for solving structural problems. Revisiting that foundation is always valuable. This time, I implemented a classic truss analysis problem using a Jupyter Notebook—building the global stiffness matrix step by step, assembling DOFs, and solving systematically. It’s a good reminder that before complex solvers and software, there was a very elegant and logical framework behind it all.

Scrolling infinite dots feels satisfying. DrawOnDot: Part 1 I’ve started working on a new project where the goal is simple but the implementation is a fun math puzzle: creating a truly infinite, scrollable grid of dots using Jetpack Compose. The challenge? You can’t actually render "infinite" objects. If you try, your memory will give up before your user does. The trick is to render only what is visible in the viewport and use the Modulo (%) operator to make the grid feel endless. How it works: 1-> State Tracking: Use detectTransformGestures to track a global Offset. 2-> The "Magic" Math: By calculating offset % dotSpacing, we find the relative starting position for the first dot on the screen. As you scroll, dots "recycle" their positions seamlessly. 3-> Performance: Instead of drawing hundreds of individual drawCircle objects, I used drawPoints. This batch-renders the coordinates directly on the GPU, keeping the scrolling buttery smooth at 60fps+.

Day 63 Rotate String Today’s problem was about checking whether one string can be obtained by rotating another. Approach 1: Brute Force (Simulation) Try every possible rotation and compare with goal using circular traversal. Approach 2: Optimized Trick If goal is a rotation of s, then it must be a substring of s + s. What I Learned: Look for patterns to optimize solutions Simple tricks can replace complex logic Always think of optimization before simulation #DSA #100DaysOfCode #CodingJourney #LeetCode #Cpp #ProblemSolving

Here's the link to the Lagrange multipliers betting game I built. You can test out your intuition on whether a point becomes a support vector or not. You can check the same under various kernels too, each with their own specific parameters. Once you have locked your bet, you'll also see the margin for the data under the specific kernel. Lastly, you see the value of the multipliers from step 0 to step 50 on a simple bar graph. The game uses SMO algorithm and KKT complementary slackness conditions to check if a vector is a support vector or not. (New) Link: https://lnkd.in/dmU2hcei

We have just released Analysis Situs 2026.1. Here are highlights: - Sphere-based thickness analysis is now open-sourced. We prototyped it a decade ago, now it's public. Good algorithms take years to mature (something the "vibe coding" people have yet to discover). - Minimal face width check for DFM and toolpath planning. Still in "test-drive" mode, looking for feedback. - Face maximization got a "protection" mode. The algorithm can silently corrupt your B-rep and pass it downstream: now it checks the UV domain before wrecking geometry. - Feature recognition "junior" by double-click: click a face and get the whole feature. Borrowed from Fusion/SpaceClaim, runs several recognizers to figure out context. - Documentation has been significantly reworked: better coverage, search, and added links to the API refman. - Many UI improvements, mostly invisible but handy: zoom-to-cursor in 3D, arrow keys for panning, less antialiasing for better FPS, progress indication in Windows taskbar, remembering last directory, and keyboard shortcuts in both console and viewer. - Fixes and improvements in data exchange slash XDE medium: SVG for UV domains, passing over STEP materials, visual diagnostics in glTF. - A bunch of fixes in the Outliner (silhouette gaps), fixed STEP import for shapes with improperly classified contours. Read more: https://lnkd.in/eBpKvQcF