Professor Norm Matloff from UC Davis is offering the community an open textbook: Programming on Parallel Machines; GPU, Multicore, Clusters and More.
Matloff describes the book as suitable for either students or professionals, as there is very little theoretical analysis of parallel algorithms. Instead he has provided extensive coverage of what he calls “wizardry” aspects–material known to experienced practitioners but generally not available in books, such as coverage of loop iteration scheduling and memory effects of storing large arrays. For the Big Data enthusiast, there is also quite a bit in there about MapReduce and Hadoop.
The book is freely available for download, subject to the Creative Commons license.