Let's create a useful book - together - for software developers interested in learning how to program effectively with SYCL.

You are invited to contribute a chapter to a contribution-based book that will focus on practical techniques for using SYCL for heterogeneous programming.

Broadly speaking – chapters should address one or both of these concepts:

1. SYCL Programming learn through examples, with explanations of use of SYCL,these examples should show results with SYCL used on multiple devices.
2. SYCL for CUDA programmerslearn through examples, with explanations of CUDA vs. SYCL,these chapters would always show NVIDIA-CUDA vs. NVIDIA-SYCL comparisons,and should also show results from other vendors as well.

Preference will be given to submissions that include examples and results utilizing Intel GPUs, either integrated GPUs or discrete GPUs, and especially both.

Submission Guidelines

A chapter in a book offers a much different opportunity to teach than a conference or journal submission - because it can be longer, include much more code listings and discussion, write in an approachable style, and discuss actual source code.

James Reinders, is also a very experienced author and editor with whom you'll enjoy working. James will work tirelessly with you, to help bring your chapter to its full potential.  James enjoys working with authors to help create wonderful chapters. You will NOT be alone!  Let's have fun together!

Please submit an idea - and work with us on schedule as part of the process.  If you have an idea that could be a chapter within the next 18 months - please submit your thoughts!

Abstracts are now being accepted on a rolling basis (the initial submission were due by February 16, 2022 and will be processed first). Submissions should be 300-600 words long. Please explain what you envision for your chapter by including (1) who will contribute, (2) a summary of what you will show/teach, (3) explain the source code that will be discussed and how much will be shared (status of source code being open source for instance), and (4) any prior presentations/publications of your topic. Feel free to just point to a published paper, and explain how you plan to expand it into a teaching chapter.

Chapters should be original work and not simultaneously submitted to another book. Expanded versions of prior work are welcome (please let us know the prior publications).

Chapters should include the following:

• Have a teaching oriented, approachable, writing style
• Discussions of both what works, and what does not (yes - share your mistakes too!)
• Source code listings, and discussions
• Option to download the full source code (completely free for at least learning purposes)
• Feel free to ask the editor (James) to help with the teaching/approachability (James likes to help edit, offer feedback and suggestions)
• Feel free to propose very rough drawing/figures or ideas for such, and ask for help developing them (in other words, artwork does not need to be camera ready - James will HELP!)

Feedback/selection on the initial batch will occur the week of February 28, 2022. Chapter submissions will be assigned one of 7 staggered assignments for delivery of the first complete draft (date will be selected together): April 26, May 31, June 21, July 19, August 23, September 27, or October 18 (later dates can be discussed as well). Feedback will be supplied within 3 weeks of receipt of chapter submission, modified (final) copy due 4 weeks later. Publication - we publish chapters on a rolling basis, and assemble books when sufficient chapters have been published.

Contributor’s agreement: Simply put the text "This work is licensed under a Creative Commons 'Attribution4.0 International' license." at the beginning of your paper.

Contributors will be acknowledged in the book. Please provide names, affliations, and a short bio for each author (put in your submission).

Submission format: Microsoft Word preferred, LaTeX accepted, other formats at the discretion of the editor (James Reinders).

Length: Chapter length should be between 14 and 30 pages (3,150-6,750 words), with generous allowance for figures, diagrams, and code listings.  A typical chapter will be 18-25 pages in final format, but some may be 30 pages long if appropriate (esp. if lots of code listings with discussions are involved).

List of Topics (must fit into at least one of these areas)

• results from SYCL code using multiple device types (e.g., CPU, GPU, FPGA) to execute the kernels
• results from SYCL code using devices from at least two different vendors (e.g., Intel, NVIDIA, AMD) for executing kernel code.
• results from a CUDA and a SYCL version of the same code. In this case, at least one comparison is on the same device. Additional results on multiple devices are welcome.
• results comparing SYCL with another accelerator method, for instance: OpenMP target offloads, OpenCL, OpenACC, HIP, CUDA, ISPC, etc. In this case, at least one comparison is on the same device. Additional results on multiple devices are welcome.

REQUIRED: Your results must be optimized similarly for each device, or well explained why not.  The goal here is to not publish results that are misleading about the underlying hardware. This should only discourage proposals that run code and the report results without investigating any surprising results.

REQUIRED: You must show the actual code, and discuss code changes, in your Chapter. Snippets of code will be shown in the Chapter, with the complete code available for download.

REQUIRED: Code (examples) discussed in the book must be freely available. In general, you must upload your code to our GitHub repository for this project, complete with   CMake (we’ll help you as needed) to compile and run your code (we’ll help – don’t worry too much about this!). Obviously, this means your code needs to be open source. You can use any open source license you desire; you need to upload the license with your code. We prefer MIT licensing.  If you are discussing code that is available in an established open-source project, we can discuss what belongs in the book repository.

Priority given to submissions which are:

• Teaching oriented: explain a challenge, discuss the best approaches to solve
• Show source code, make source code available (open source)
• Preference will be given to submissions that include examples and results utilizing Intel GPUs, either integrated GPUs or discrete GPUs, and especially both.
• Do not assume a hardware background

Publication

Each chapter will be made available as a free PDF download, and via print-on-demand, and will be assigned an ISBN number.  Books (of 15+ chapters) will also be available freely as PDFs, and via print-on-demand, and will have ISBN number.

Contact

You may email us with questions.

Please submit sooner rather than later... no obligation, it just starts a conversation with us.