I attended Mapbox Locate recently and after Lauren Budorick's presentation on mapbox-gl I wondered if the power of GL could not be leveraged for performing spatial queries on geometric data that have already been loaded into a browser, as an alternative to turf.js on the front end. Is this something you've pondered, and/or do you see any viability in pursuing such a thing?
Engineer at Mapbox, creator of Leaflet, open source enthusiast, algorithms geek, speaker, singer-songwriter (Obiymy Doschu), father of twin girls, Ukrainian
WebGL is great for rendering large amounts of data, but is very difficult to use for actual processing and analysis. There are many research papers with geometric algorithms tailored for the GPU for massive parallelization, but pretty much no real implementations, and those algorithms use advanced APIs such as compute shaders, which are currently not available for WebGL and won't be for a while. Khronos is working on such an extension for the upcoming WebGL 2.0, but it will take a long time.
For further actions, you may consider blocking this person and/or reporting abuse
We're a place where coders share, stay up-to-date and grow their careers.
I attended Mapbox Locate recently and after Lauren Budorick's presentation on mapbox-gl I wondered if the power of GL could not be leveraged for performing spatial queries on geometric data that have already been loaded into a browser, as an alternative to turf.js on the front end. Is this something you've pondered, and/or do you see any viability in pursuing such a thing?
WebGL is great for rendering large amounts of data, but is very difficult to use for actual processing and analysis. There are many research papers with geometric algorithms tailored for the GPU for massive parallelization, but pretty much no real implementations, and those algorithms use advanced APIs such as compute shaders, which are currently not available for WebGL and won't be for a while. Khronos is working on such an extension for the upcoming WebGL 2.0, but it will take a long time.