We present below some basic tests on the performance of the ddCOSMO-algorithm. The purpose of these tests is not to give an exhaustive insight of the performance, which can be found in the literature indicated here, but is thought as an apetizer.
Figure 1: Elapsed time in seconds as a function of the number of atoms for the solution to the direct (X) and adjoint (S) ddCOSMO equations and for the ddCOSMO part of the forces (F)
Figure 2: Performance gain as a function of the number of shared-memory cores for the solution to the direct (X) and adjoint (S) ddCOSMO equations and for the ddCOSMO part of the forces (F)
The computations were carried out on a 2*Xeon E5firstname.lastname@example.orgGHz (16 cores), 64GB DDR3@1.6GHz cluster node. The ddCOSMO module was compiled with the Intel Fortran Compiler 14.0.1, using aggressive optimization (-O3 flag).