Skip to content

Book previews and readers

Checked and downloaded on 2026-09-23. These are public previews, not full ebooks. Use the publisher links below for the full editions with your own purchase, subscription, or institutional access.

Book Local download (not in Git) What the file contains Read online
Harshwardhan Fartale, GPU Programming with Triton (Manning) Chapter 1 preview (HTML): downloads/gpu-programming-with-triton-chapter-1-preview.html A short opening excerpt and chapter outline, not the complete chapter. The saved publisher page may need internet access for styling and images. Public chapter brief · LiveBook reader · Publisher / MEAP access
Chris Fregly, AI Systems Performance Engineering (O'Reilly) Table of contents (PDF, 14 pages): downloads/ai-systems-performance-engineering-toc.pdf The publisher's contents list only; it does not contain the chapters. Provided by Shroff/O'Reilly for its Indian edition, ISBN 9789368087298. O'Reilly reader and access · Public PDF download · Download's publisher page

Manning lists the Triton book as early access, with 4 of 11 chapters available and publication estimated for spring 2027. A complete finished edition is not available at the time of this check. No freely downloadable full edition of either title was found on the checked publisher pages.

How to use them with this repo

  1. Open Manning's brief to see the scope, then start the repo's orientation and indexing lesson.
  2. Use the O'Reilly contents PDF to locate the GPU/CUDA foundations and memory topics in chapters 6-7. Chapters 13-14 cover PyTorch profiling and the Triton compiler path. Reading those chapters requires access to the full book.
  3. Use the freely readable upstream tutorials alongside our exercises while deciding whether to obtain a book.

The original downloaded files are unchanged. Their URLs, sizes, and SHA-256 checksums are recorded in sources.json. The PDF was checked for its title, 14-page count, readable contents, and rendering.

Downloads are stored locally under downloads/ and ignored by Git so the portfolio does not redistribute book material. On a fresh clone, use the online links above. Licensed personal ebook files can also be kept in that ignored folder.

Back to all references or the course.