Tablets of parallel computing - what do you recommend?

Question

Books

What, in your opinion, are really worthwhile books / courses on this topic? Indicative level: any, from the basics to the very peaks - I will re-read / revise everything.
From what I found - The Art of Concurrency (didn't see it in the Russian version, so I'm not sure about the Russian-language name) - what is your opinion about it?

Answer 1

start from here .

Answer 2

hpcc.unn.ru/?dir=18 look here.

Answer 3

www.amazon.com/gp/product/0123705916

Answer 4

Parallelism is too general a concept to be able to answer something specific to such a vague question. If you need to be able to program scientific applications on clusters, then start learning MPI. If you need the same scientific and financial, but on mainframes or workstations, then study something on shared memory: OpenMP, Cilk Plus, Intel TBB or abstractions of working with threads from QT or boost.
But it's all the "soft way". You can use hardcore TCP IP instead of MPI, and pthreads instead of shared memory libraries (for linux). And if you also plan to use Nvidia / AMD graphics accelerators, then additionally study CUDA (for Nvidia) or OpenCL (universal). Moreover, parallelization in the sense of accelerators and processors will have to be very different.
But all this taken together may turn out to be completely unnecessary if the goal is to use parallelism within a language or a specific platform. C#, Java have internal threading controls. To whom few can add Scala, Node.js - there is very advanced flow control. And this knowledge will not intersect with the first two paragraphs at all.
I have used 90% of the listed technologies. There is no "best" and cannot be - it all depends on specific goals and needs. So just start with the basics of what concurrent programming is. Further - according to needs.