Parallel computing on Android, a niche topic
Alberto Marchetti
Senior DevOps Engineer (Full Stack Developer in my previous life!)
Parallel computing is, nowadays, a common topic: there exist some well-known libraries, like CUDA, AMD APP and OpenCL, which let developers enhance softwares' performance.
There is, however, a kind of parallel computing that is still not widespread: mobile-based parallel computing. Both Google and Apple provide frameworks to perform parallel computations on their operating systems: RenderScript by Google and Metal by Apple.
Whoever wants to approach parallel computing on mobile phones has an easy and immediate choice: RenderScript.
RenderScript is a really easy-to-approach parallel computing framework, having both basic and advanced functions. Despite having few online recent documentation, the book "RenderScript: parallel computing on Android, the easy way", can let developers learn its basics in a blink of an eye.
Learning to use RenderScript basics can provide a certain degree of specialization to developers, which can be easily appreciated by software development companies, being it today a niche topic, not so well known and understood.
RenderScript can be used to perform certain parallel computing operations that could be applied only to a mobile-based market, like real-time mobile-camera video processing for statistical analysis, as well as real-time high-frequency mobile-microphone sound analysis.
Give it a try, and you'll find out a beautiful framework, which gets frequently updated with new features and can help you in ways you cannot even imagine right now, which you can discover as soon as you start thinking using RenderScript-developers point of view.
Alberto Marchetti
PhD(Wavelets).MS.BE.(ME,EE), Owner CRD P/L, Father of the Australian Worsted Comb Fibre processing Technology
7 年I have tried it using an Android Linux but experiencing on two AT300 Toshiba tablets, with a common HDD on a local dedicated network. Linux lens it easier to do that then Windows server - using Win HPC 2008. Linux is way to go, I am trying to make it portable by using a WiFi NAS server HDD and two tablets - I can expand that over more Android type tablets. Just imagine several quadcore Tegra processors doing low power super computing....I am 60 and I am impressed what has transpired over my working life as an engineer in terms of computing performance. Keep up the good work - aim is to link all devices in a computational network that one has in their home as we all have more then one computer.