Felipe de Alcântara Tomé
Felipe de Alcântara Tomé
This PR wraps up the base version of the DArray for MPI, including randomized allocations, ones, zeros, distribute without specified partition dimensions, and all the variations of reduce (sum, prod,...
Why cant we get arbitrary precision on the besselj functions?
This PR concerns the implementation of parallel Triangular and Trapezoidal wrappers. A parallel Tile QR implementation, and undefined initializers for DArrays. This gives ways for "zero-memory" allocation of trapezoidal and...
In the fda/MPI2 branch the chunks cannot be waited on if their internal parameters are MPI related
MPI suite and base implementations such as processors and messaging with P2P MPI directives that fit into both the scheduler logic and datadeps scheme.
This includes helpers for creating identity DMatrix, undef DArray, QR implementations and minor testing
What would be necessary to exploit the Hwloc numa distance feature (that is available in the wrapper already)? I'm interested in implementing it and wanted to follow the templates/models the...