Concurrency and Parallel Computing in JavaScript
summary
Overview of concrete artifacts and limits: River Trail proposes parallel array methods like mapPar reducePar scan scatter filter and has a River Trail prototype on GitHub that targets multi core CPUs and GPUs and interoperates with typed objects the proposal builds on typed objects and the uploader names River Trail and Intel and Mozilla involvement promises and Web Workers are recommended for error propagation and long tasks the talk shows SIMD float32x4 intrinsics as a concrete optimization and notes Firefox Nightly implements the array part and work is ongoing
AI-written summary · about
chapters (10)
- 0:00Introduction to concepts
- 1:31Concurrency on the web
- 4:20Promises and web workers
- 4:46The case for parallel computing
- 5:49Promises for asynchronous flow
- 17:46Web workers for performance
- 21:32Parallel computing strategies
- 24:13SIMD programming in JavaScript
- 32:11Parallel JavaScript with River Trail
- 39:52Conclusion
related talks
23mParallel JavaScript
25mGetting Over Moore's Law: Parallelization Using JavaScript in the Browser
24mShared Memory & Parallel Programming with JavaScript
33mParallelism experiments in JavaScript
47mLOLJS
32mMulticore JavaScript: Past, Present and Future
33mAsync patterns to scale your multicore JavaScript elegantly.
29mParallel computing in ReactJS: Provide a pleasant user experience