concurrency - Why do books on concurrent programming always ignore data parallelism? -


In the last few years there has been a significant change towards data-parallel programming through systems such as OpenCL and CUDA, and so far Books were also published in the last six months, never even the subject of data-parallel programming.

It is not suitable for every problem, but it seems that there is an important difference that is not being addressed. First of all, I would suggest that concurrent programming is not a parallel equivalent to parallel programming. Concurrent programming is about creating applications with loose-coupled tasks. For example, a dialogue window could have interaction with each control executed as a separate function. On the other hand, parallel programming is about broadcasting clearly about a computational task solution more than a piece of hardware, essentially always for any kind of performance reasons (note: even a lot Low RAM is an alternative cause when there is optional swapping.

So, I have to ask in return: are the books you are talking about? Do they concurrently About gramming (I have some of these, there are so many interesting principles), or parallel programming about it?

If they are actually about parallel programming, then I have some comments

  • The CUDA is a fast moving target, and since its release, the book written about today will be made in print by half as obsolete.
  • OpenCL's mother A year was released a year ago.In the last 8 months, stable implementation came, there is not enough time to get a written book yet, only modify and publish.
  • Open NMP at least Some parallel programming is covered in textbooks which I have used since version 2 (v3 was released now), it was basically about data parallel programming.

Comments

Popular posts from this blog

c# - How to capture HTTP packet with SharpPcap -

php - Multiple Select with Explode: only returns the word "Array" -

jquery - SimpleModal Confirm fails to submit form -