TABLE OF CONTENTS / Introduction / Current State of the Art / Cutting and Machining / Ship Stability / Sea Ice Motion / Astrophysics / The Near Future of Parallel Computing / Afterword /
/ 6 Figures /
The complexity of modern computer codes today, together with rapid developments in both software and hardware, practically guarantee the codes' obsolescence tomorrow, when their original developers have gone. The Livermore Laboratory's Dyna3d and ParaDyn codes, for example, are hundreds of thousands of lines long, with complex physical models. The complexity makes the computation of total momentum and energy onerous, so that these important diagnostics are missing in the current software. Because the state-of-the-art software is growing constantly more complicated, assessing the state of the art is very much an attempt to track a moving target.
SPAM has always been an excellent computational technique for following flows of liquids or gases under positive pressure. The treatment of solids or fluids under tension, particularly for long times, requires special techniques. Explicit surface tension, velocity averaging, and density-gradient and invariant-curvature potentials are examples of such techniques mentioned in this book. Solids present an additional difficulty: angular momentum tends to dissipate artificially. Perhaps a technique assigning individual particle angular momenta ( so that the "particles" actually spin, to acquire an angular momenta commensurate with their size and rotation rate ) to individual particles would be useful.
In much published research ( see the examples cited in this Chapter, all of which were found on the internet...
