adda
adda copied to clipboard
Shared-memory execution
Optimize ADDA for shared memory hardware (e.g. multi-core processors),
using standard MPI functions and/or OpenMP directives.
Original issue reported on code.google.com by yurkin
on 24 Dec 2008 at 7:13
It should be trivial to employ multi-threaded version of FFTW - just a matter
of a couple of initialization calls. The only challenge with that is rigorous
compilation (Makefiles, etc.) to work on all OS with different thread models.
Original comment by yurkin
on 23 Jan 2012 at 2:27
The following may be marginally relevant - http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.116.2931&rep=rep1&type=pdf
Another option is to use MPI shared memory model (in MPI 3). It will allow efficient execution on both multi-core processors and multi-node clusters