DeepRead
Cover of Programming Massively Parallel Processors

Programming Massively Parallel Processors

David Kirk, Wen-Mei W. Hwu · 2010

Programming and computer science › Operating systems and networking

#2,906 most recommended

About the book

A textbook on writing parallel programs for graphics processing units, using NVIDIA's CUDA model as the main vehicle. It covers the GPU execution model, thread organization, the memory hierarchy, and the algorithmic patterns that map well to massively parallel hardware, such as reduction, scan, histogram, and stencil computations. Writing fast parallel code requires understanding the hardware, not just the syntax of a parallel language. The authors contend that performance is governed mainly by memory access patterns, data locality, and how work is divided among threads, so programmers must reason about these explicitly. They present parallel decomposition as a problem-solving discipline that applies beyond any single platform, and they use worked examples to show how a naive kernel can be restructured for much higher throughput.

Written by Claude from what it knows of this book.

What Hacker News says 53% recommend · 0% critical

02.5520112015202020262024 peak · 5
This bookAverage book in Operating systems and networking

In the commenters’ own words

What Open Library readers say

No Open Library reader has rated it yet.

  • 2 want to read
  • 1 reading
  • 0 have read

Open Library’s page for the book

What critics say, via Wikipedia

No Wikipedia article about this book yet, so no critics to quote.

Reddit communities — coming later

The threads it came up in

Details

First published
2010
ISBN
9780323912310
Editions known
6 at Open Library
Open Library
OL25666151W

Computer architectureMultiprocessorsParallel processing (electronic computers)Parallel programming (computer science)

Explore in the mosaic

Read next: its neighbours on the mosaic