Blog Entries

...

Dependence of efficiency on the number of cores is a very important characteristic of scalability solver based on which one can predict the scaling limit. For simplicity we consider only one option for the solver, using the asynchronous version of the MPI exchanges Isend() + Irecv(). It is seen that the dependence of the efficiency of parallelization of the solver is linear or nearly linear, with a gradual drop in effectivity with increasing number of cores used
...

The other day a fellow forum user asked me for help about a problem dealing with the specific configuration of a case for running in a cluster. The specific configuration was simple: each slave machine of the cluster has it's own independent storage address on a similar path.
Below is an edited version of the reply I sent him.


OK, first detail - In the following file you have a (sort-of) well documented "decomposeParDict": ...

...use a blocking call Sendrecv() instead of a pair of Send() + Recv() slightly improves the performance of the code but does not result to improved scalability
...

Let us consider in more detail how works domain decomposition in the simplest 1D case.
The computational domain is divided into several subdomains along one (usually the longest) dimension. The number of such subdomains usually corresponds to the number cores involved. The main data exchanges take place between neighboring subdomains. The size of transmitted data between adjacent subdomains in the general case depends on many factors (number of equations in the system, the size of the computational...