"My" book (others helped too!) |
Our book got published end of last year, but due to shipping problems and holidays I only got my hard copy a little over a week ago. This project started in May 2017 with Yutaka Ishikawa of RIKEN proposing that the four of us undertake such a project. We began by outlining the structure of the book and thinking about whom to invite to participate. We wanted to cover the last three decades in operating systems for supercomputers and hoped for contributions from all major players during that time. Forty authors from three continents joined us in the end. One glaring omission are the recent successes of the Chinese in the Top500 list. Despite trying quite persistently we were unable to get a chapter on that. There are other omissions too, for example the K40 project, where the book remains quiet. But, all in all, it is a nice history and lessons learned that covers a large number of projects.
Once we had the chapter list we worked with the authors and Springer Verlag to to give the book a logical flow, consistency, and a cohesive feel. This is not a bunch of isolated chapters thrown together. The book is organized into three major parts, each roughly covering one decade between the 90s and now. Each part has an introduction to explain what motivated the work done during that time, and the influence things like the arrival of MPI and Linux, and the move to commodity components, had on the projects described in the book. There are a common bibliography, index, and list of acronyms that cover the entire book.
It was a lot of work and it was fun to work with the people who contributed to the book. I am very proud of the end result. Not only because I think it is a good book, but also because it covers many of the projects I worked on over the last, nearly thirty years, and many of my friends and colleagues from that time contributed.
Some Easter eggs for you. The SUNMOS and Puma section has pictures I drew nearly thirty years ago using Framemaker. Current versions of that product can still open those old files and convert them. That is not the case for some other tools. I have PowerPoint and Word files that can no longer be viewed. One of them is a depiction of a cluster I drew for the first SOS workshop. There are jpeg versions still in circulation. For the book I created a new image. Nowadays, supercomputers have impressive images on their front panels. (Back then we had to do with blinking lights.) So, I superimposed an image of Mt. Hood, tallest peak near my new home, onto the front of my new drawing. I took that picture on a hike with my daughter. The book cover picture is also mine. I took it when we got to see Oakforest PACS in Japan.
The book topic is pretty specialized but it should still be an interesting read for people interested in systems software and supercomputers. And No, you don't have to buy one to support me or the authors. Our pay was a hard copy of the book (two for editors), and that's all we'll ever get. But it is nice to be able to go to Amazon or your favorite online bookshop, type in your name, and see the book pop up.