[mpich] MPICH primary repository branch, master, updated. v3.1b1-101-gc580679
This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "MPICH primary repository". The branch, master has been updated via c580679c9a583486ad321c885aaeb13351018716 (commit) from 011cd3dfd015608babb640b35266a792231e8833 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- http://git.mpich.org/mpich.git/commitdiff/c580679c9a583486ad321c885aaeb13351... commit c580679c9a583486ad321c885aaeb13351018716 Author: William Gropp <[email protected]> Date: Sun Oct 20 13:42:13 2013 -0500 Ensure MPIR_COUNT_ALIGNMENT set In some cases, MPIR_COUNT_ALIGNMENT was not being defined. This fixes that case, which came up when trying the -feature=fint=on option for the Fortran 77 code. diff --git a/configure.ac b/configure.ac index 8d11040..e50d434 100644 --- a/configure.ac +++ b/configure.ac @@ -5843,12 +5843,12 @@ if test "$enable_f77" = yes -a -z "$MPI_STATUS_SIZE" ; then # size and then tell Fortran to enforce maximal alignment. # A more sophisticated test would determine the required alignment, # rather than assuming alignment to maximum size + MPIR_COUNT_ALIGNMENT=0 if test $MPI_SIZEOF_COUNT -gt $pac_cv_f77_sizeof_integer ; then # Alignment test looks at the low address bits, as long # as the sizeof(MPI_Count) is 2^j. # If we can't tell, pick a value that forces the Fortran # wrappers to use a local copy. - MPIR_COUNT_ALIGNMENT=0 case $MPI_SIZEOF_COUNT in 4) MPIR_COUNT_ALIGNMENT="0x3";; 8) MPIR_COUNT_ALIGNMENT="0x7";; @@ -5856,8 +5856,6 @@ if test "$enable_f77" = yes -a -z "$MPI_STATUS_SIZE" ; then *) MPIR_COUNT_ALIGNMENT=-1 ;; esac - AC_DEFINE_UNQUOTED([MPIR_COUNT_ALIGNMENT],[$MPIR_COUNT_ALIGNMENT],[Value that anded with an address will give 0 if the address is properly aligned for MPI_Count]) - # Compare the size of MPI_Status to the sizes of the individual # elements. If the sizeof(MPI_Status) > sum of the sizes, then # determine if simple padding will work. @@ -5929,7 +5927,10 @@ _EOF AC_MSG_RESULT([no]) fi fi -fi + # Define the alignment of count regardless of whether sizeof(MPI_Count) + # is bigger than a Fortran integer + AC_DEFINE_UNQUOTED([MPIR_COUNT_ALIGNMENT],[$MPIR_COUNT_ALIGNMENT],[Value that anded with an address will give 0 if the address is properly aligned for MPI_Count]) +fi # enable_f77 and mpi_status_size not set AC_SUBST([MPI_STATUS_SIZE]) MPIF_STATUS_SIZE=$MPI_STATUS_SIZE AC_DEFINE_UNQUOTED([MPIF_STATUS_SIZE],[$MPIF_STATUS_SIZE],[Size of an MPI_STATUS, in Fortran, in Fortran integers]) ----------------------------------------------------------------------- Summary of changes: configure.ac | 9 +++++---- 1 files changed, 5 insertions(+), 4 deletions(-) hooks/post-receive -- MPICH primary repository
participants (1)
-
noreply@mpich.org