Caro-COOPS.org Discussion Board

jcothran · Joined: 24 Feb 2025 Posts: 176 Location: Columbia, SC

The computational science behind model simulation and parallelization is a broad problem with application across a variety of computationally intense domains. I thought the following article(a general read) and white paper(a specific technical read) describe some interesting approaches/simplifications to reducing the model computational load. The first talks about changing the load given a beginning number of resolution points considered from an exponential function to a logarithmic function so that a higher resolution model does not make things computationally possible.

Simulating the Whole Universe
http://radio.weblogs.com/0105910/2004/09/04.html

Fully threaded tree algorithms for massively parallel computations
http://www.hpcmo.hpc.mil/Htdocs/UGC/UGC00/paper/alexei_khokhlov_paper.pdf