The Art of Molecular Dynamics Simulation

Strengths: Excellent selection of topics, with good references and enough discussion of the underlying scientific theory. Thoughtful and well-written. DO NOT USE MERELY AS A COOKBOOK! Weaknesses: Antiquated code, basically FORTRAN with a C veneer. Although the code works as is and is blazingly fast, it's not usable as a basis for further code development -- readers will want to recast the ideas and algorithms in their own up-to-date C code. For example, his arrays start at 1; the only data structure is the array (ugh!), there are no structures. Things get hard to understand when he starts packing multiple array indices into a single composite index, then unpacks it (see, e.g., the code on autocorrelation functions, Chap. 5). However, the awkward code is a blessing in disguise -- write your own and you'll learn an incredible amount! This book gives you all the conceptual tools to do this, and you can use the code supplied (available from the publisher) as a benchmark and check. Actually, there are a few bugs in the code, but considering its complexity (much of which is unavoidable), it is remarkably reliable. In general, the code is optimized for efficiency, not for ease of maintenance or generalization. The author deserves any amount of thanks for writing this unique and indispensable book, which is clearly the fruit of many years of practical computing experience in a 70's, early 80's UNIX environment. The code has not aged well, but the ideas it embodies are timeless.

I'm not sure I'm that reader. The ideal reader is wholly at home with good bits of analysis - you'll be comfortable with Lagrange multipliers, LU decomposition, quaternion representation of oriented values, and a fair bit else. You won't need the basics of classical physics, including Hamiltonians and a few projections of the Navier-Stokes equations down into tractable form. Non- and in-equilibrium dynamics, smooth and discrete representations, plus their mixed forms - you're prepared to build up from these elements. You, the ideal reader, are not much of a programmer, though. The book built very much around extensive C sample code. I use the term "C" in some strict syntactic sense, though. The general style reminds me of FORTRAN, but the global data declarations are in the style of pre-1980 BASIC. This book is aimed at the odd mix of capable numerical analyst (which I'm not) and chemist (which I'm not), but naive software developer (which I'm not). The strength of this book is in its code samples. The author explictly invites modification and extension. The only sane way to modify this code is to understand it thoroughly, and to understand the underlying chemistry, physics, and numerical models. Maybe this book works well for some people, people that I admire immensely. I just wish it worked better for me. //wiredweird

This book is in some ways a reasonable introduction to MD. It's the only book I know that actually shows the derivatives of energy functions necessary for writing efficient MD code. Unfortunately it does so in a hard to follow way. This is typical of the rest of the text. The example C code is hideous both in format and style. Missing from the book are essential formulae for calculating macroscopic properties from MD simulations. It is probably the best practical guide to writing MD code, but that isn't saying much.

The code examples are in C, but the author tries to write it "Fortran style". Basically it seems almost as if the code was converted from Fortran code to C by f2c (a GNU translation program). It really seems as if the author's background is in Fortran and hasn't really grasped the fundamental C approach to writing code. However, if you can manage to decipher the code, the technical information is very useful. Most molecular dynamics books still seem to be directed at the Fortran audience, so this is probably as close as you will come to a book for the C audience.

This is a great book on molecular dynamics. It covers lot's of interesing topics. Unfortunately most of the info is not explicitly written, but hidden in the included code. If you already know how to program in C and take the time to examine and analyize the programs you can get some really good code.

This book constitutes a good introduction to the methods of Molecular Dynamics. Besides clear descriptions of the methods and algorithms, it includes C example implementations commented on the book to illustrate the points. It is almost an introducory 'recipe' book. For advanced tips, Frenkel-Smit will give good lead, but at the cost of less readability and Fortran (pseudo)code. Certainly, the underlying philosophy of the code may seem awkward initially, but it is not a matter of "C design", rather of coding philosophy and program design. Once you get to understand it, it becomes easy and "natural". And what's even more important: it reflects the program design philosophy in use in the Molecular Modelling field, hence getting used to it will help you understand most existing production code. The electronic version of the C code (available at the book's web site) is a mesh though: it lacks comments and makes heavy use of conditional compilation to render the examples from a single, convoluted source. The ingenious programmer won't be misled by this however: a simple step with the C preprocessor will distil the actual code. Not to mean it is not inconvenient anyway (sic). Certainly, it would have been a lot better if the authors had separated, cleaned and commented the code for the examples in the electronic version. Besides, the book comes with additional, not too difficult programming assignments that make for extra fun and deeper understanding. Once you are done with this book, you'll be in a position to really appreciate and use Frenkel-Smit and other advanced texts.

This is a good and honest book on the subject, it needs a second edition in pure C language and not the 'f2c-like' used. Many techiques like linked lists are simple and clear in C but confuse when translated to Fortran (77 or the new 90/95 variants)

The code seems to me unclear ,unelegant and slow. Code of the book on the web is unnecessarily complex and hard to understand. This is a book of simplistic recipes without formal explanation. So is not usable to learn or understand. Obviously because it was not written in C originally and translation to C was made in a rush. If you are learning buy Statistical Mechanics by Rowley , Understanding Molecular Simulation by Frenkel or Computer Simulation of liquids by Allen.