|
From: Joseph W. <joe...@gm...> - 2013-10-17 12:15:16
|
Just a heads up. I'm working on getting QuantLib working with parallel computing. I've got the system built with OpenMP, and I'm going to be playing with getting it working with OpenMPI afterwards. I'm going to start with the "obvious" parallel areas (i.e. parallelize monte carlo paths) but if someone has ideas on non-obvious places to look at parallel coding let me know. One non-obvious place I can think of is the tridiagonalization routine where you have to use a different algorithm to get any parallel speedup. |