This is a bug-fix release that only
- adds a file missing from the distribution archive (but is only needed when time-stamps get changed)
- fixes a problem when ifort or icc was used with the -diag-disable option having an argument as next command-line argument instead of separated by an equals sign
- fixes needed if autoreconf is performed with automake 1.16
The new YAXT release includes
- support for asynchronous exchanges (like MPI_Isend) and
- fixes for MPI implementations where MPI_Pack_size returns an actual upper bound and not the exact amount that MPI_Pack advances position by
- document the consequences segmented memory architectures might have
The following improvements have been added to YAXT:
- User-defined redists via the
xt_redist_single_array_base_new function
- Lots of minor cleanup and bug-fixing
This is mostly a bug-fix release but also includes
- the option to change the RMA synchronization method of distributed multi-arrays,
- fixes to enable builds with g95,
- Kahan summation is right out-of-the-box on more systems (i.e. one less source of twiddling compiler flags).
The following improvements have been added to YAXT:
- Improved g95 support.
- A memory free bug when concatenating empty redists.
- A minor bug fix for stripe overlap computation when stripe starts are far apart and stride is large.
- New constructor xt_redist_repeat_asym_new allows for different displacements in source and destination arrays.
- Various minor fixes to tests.
The following improvements have been made to YAXT:
- The C part is now available as a separate library so that the differences in linking between Fortran and C can properly be accounted for.
- As a consequence of the above, friendlier generic interfaces are available for Fortran and mpif.h is no longer needed for building YAXT.
- Intercommunicators are supported for constructing Xmaps (destination indices can only be satisified from sources in the remote group) for both,
xt_xmap_all2all_new and xt_xmap_dist_dir_new.
- Correctness and performance have been improved for index stripe lists (
xt_idxstripes_new) and the distributed-directory based Xmap construction.
- Various internal and test-suite improvements.
A new version of PPM which
- much expands C support of functionality previously only available in Fortran
- adds the distributed multi-array data structure especially to the C part
- full configure-time support for libtool
- minor fixes for various platforms and added support for BlueGene/Q
In summary the changes are:
- Better support for BlueGene/Q and cross-compilation.
- Copy constructors for xmaps and redists.
- Support for overlapping redist repeat constituents.
- Stricter declaration of some API functions (added const).
- 2 edge cases in stripe overlap got fixed.
- The tests got improved in numerous little places.
- Internal memory allocations have slightly less overhead.
This release fixes the prevention of unnecessary duplication of MPI communicators from 0.4.2 to finally work in all cases.