wllama v4.0 ditches wasm mem64 for mem32 (up to 22% faster), 2.8x speedup on Chrome
ngxson · x · 2026-11-23
wllama v4.0.0 is out, bringing up to 2.8x faster inference on Chrome, video input, full-duplex realtime audio, and hybrid CPU + WebGPU support on Chrome and Firefox for running bigger models.
Author ngxson explains a counterintuitive design choice: v4 uses wasm mem32 instead of mem64 because mem32 is up to 22% faster. To work around mem32's 4GB heap limit, the wasm runtime is split into multiple instances linked together via RPC — building on rgerganov's GGML RPC backend, applied in an unexpected place.
Related event: wllama v4.0.0 Boosts In-Browser LLM Inference Up to 2.8x(4 posts)→
More from coding & agent
- Replit agents speed up NumPy 18.9x — but only Claude's team scales — amasad · 2026-11-23
- A better yardstick for 10x engineers: how much of their work gets reused — vboykis · 2026-11-23
- Microsoft rebrands Azure AI Search as Foundry IQ, a knowledge layer for agents — adnan_hashmi · 2026-11-23
- Gemini agent builds rigged Optimus robot in Blender with 145 parts from photos — majidmanzarpour · 2026-11-23
- Dev levels up Neon-built escape room game with an epic AI-generated intro video — jacalulu · 2026-11-23
- ChatGPT Edge extension jumps from 220k to 1.4M users in 30 days, with a long permissions list — UnableCar9092 · 2026-11-23