`shard.allgather` concatenates along a specified gather-axis. However,
`mpi.allgather` always concatenates along the first dimension and there
is no MPI operation that allows gathering along an arbitrary axis.
Hence, if gather-axis!=0, we need to create a temporary buffer where we
gather along the first dimension and then copy from that buffer to the
final output along the specified gather-axis. This is not ideal by far.
Along the way also
- fixing computation of memref size in mpitollvm
- adding a simple canonicalization pattern for comm_size for easier
debugging
- adding more tests
This is replacing #125361
- communicator is mandatory
- new mpi.comm_world
- new mp.comm_split
- lowering and test
---------
Co-authored-by: Sergio Sánchez Ramírez <sergio.sanchez.ramirez+git@bsc.es>
This patch fixes:
include/c++/14/bits/unique_ptr.h:93:2: error: delete called on
non-final '(anonymous namespace)::MPICHImplTraits' that has virtual
functions but non-virtual destructor
[-Werror,-Wdelete-non-abstract-non-virtual-dtor]
* The first set of patterns to convert the MPI dialect to LLVM.
* Further conversion pattern will be added in future PRs.
* Supports MPICH compatible MPI implementations and openMPI, selectable through DLTI attribute on module
---------
Co-authored-by: Anton Lydike <me@antonlydike.de>
Co-authored-by: Christian Ulmann <christianulmann@gmail.com>
