E81 Concurrency and parallelism with Paul Butcher
ClojureStream Podcast
00:00
The Difference Between Reducers and Transducers?
Data parallelism is a spectacularly powerful programming technique if you've got the kind of problem for which it's applicable. Image processing was the original problem that managed to leverage data parallelism effectively and it's now branched out into many other areas but image processing was where it started. Instead of talking about an ATEX kind of performance increase that you might get from using multiple cores we are talking about 32 or 100 times performance gains. And even if you do not find yourself writing this kind of data parallel code directly you will almost certainly be using it in machine learning libraries.
Transcript
Play full episode
