REVIEW - Matrix Computations


Matrix Computations


Gene H. Golub, Charles F. Val Loan


Johns Hopkins University Press (2013)




Paul Floyd


March 2023



Verdict: Recommended

I’ll start off with a little niggle unrelated to the content. This book has a sort of matte cover (and without a dust jacket) so it soon started to look a bit grubby. I know that it isn’t an ornament, but I do prefer a glossy cover that I can give a wipe.

I’ve read a few books on linear algebra. Usually they are more mathematical with little or no references to code or algorithms. This book, whilst it’s not a full blown software book, does contain a lot of outlines of algorithms. Generally this is with a Matlab-like syntax, but it should be clear enough to allow most programmers to implement in any language.

I won’t go into detail of all of the 12 chapters, but I’d say that it covers everything that you are likely to encounter at a beginner to intermediate level. In short that covers fundamentals, direct solvers, iterative solvers, dense and sparse methods.

There’s a good bibliography at the start of the book (with a list of the 27 most important). Another thing that I liked was the page in the preface that describes the ‘Common Notation’. At one point I mixed up in my head what ran() is, and looking back to the page I could quickly see that it is range and not rank().

If you are looking for a book on linear algebra then I suggest that you double check the contents and ensure that there is not too much for your needs. For instance, if you just want to learn about direct solvers then a ‘bible’ like this might be overkill.