WebGPU
6 mentions across all digests
WebGPU is a web standard for GPU-accelerated computing in browsers, enabling on-device LLM inference via extensions like Gemma Gem and studied for dispatch overhead characteristics across four GPU vendors and multiple browsers.
Lumina – a statically typed web-native language for JavaScript and WASM
Statically-typed Lumina language bridges JavaScript and WebAssembly with Hindley-Milner type inference, targeting reactive UIs and WebGPU workloads in a single type-safe system.
Myth Engine Architecture: Building an SSA-Based Declarative Render Graph
Myth Engine implements a declarative render graph architecture based on Static Single Assignment (SSA) to manage GPU resource synchronization and state complexity in modern graphics APIs. The article documents three a...
Show HN: Gemma Gem – AI model embedded in a browser – no API keys, no cloud
Google's Gemma 4 model runs natively in Chrome via WebGPU with full agent capabilities (DOM interaction, form filling, JS execution) and 500MB/1.5GB variants—eliminating API keys and cloud dependency while preserving privacy.
Characterizing WebGPU Dispatch Overhead for LLM Inference Across Four GPU Vendors, Three Backends, and Three Browsers
WebGPU dispatch overhead (24–71 μs) is the true LLM inference bottleneck in browsers, not compute—torch-webgpu provides a PyTorch backend while revealing prior benchmarks massively overestimated costs by ~20×.
Transformers.js v4 Preview: Now Available on NPM!