== Highlights ==
1. New mpifort wrapper compiler: you can utilize mpif.h, use mpi, and use mpi_f08 through this one wrapper compiler
1. mpif77 and mpif90 still exist, but are sym links to mpifort and may be removed in a future release
1. The mpi module has been re-implemented and is significantly "mo' bettah"
1. The mpi_f08 module offers many, many improvements over mpif.h and the mpi module
This stuff is coming from a VERY long-lived mercurial branch (3 years!); it'll almost certainly take a few SVN commits and a bunch of testing before I get it correctly committed to the SVN trunk.
== More details ==
Craig Rasmussen and I have been working with the MPI-3 Fortran WG and Fortran J3 committees for a long, long time to make a prototype MPI-3 Fortran bindings implementation. We think we're at a stable enough state to bring this stuff back to the trunk, with the goal of including it in OMPI v1.7.
Special thanks go out to everyone who has been incredibly patient and helpful to us in this journey:
* Rolf Rabenseifner/HLRS (mastermind/genius behind the entire MPI-3 Fortran effort)
* The Fortran J3 committee
* Tobias Burnus/gfortran
* Tony !Goetz/Absoft
* Terry !Donte/Oracle
* ...and probably others whom I'm forgetting :-(
There's still opportunities for optimization in the mpi_f08 implementation, but by and large, it is as far along as it can be until Fortran compilers start implementing the new F08 dimension(..) syntax.
Note that gfortran is currently unsupported for the mpi_f08 module and the new mpi module. gfortran users will a) fall back to the same mpi module implementation that is in OMPI v1.5.x, and b) not get the new mpi_f08 module. The gfortran maintainers are actively working hard to add the necessary features to support both the new mpi_f08 module and the new mpi module implementations. This will take some time.
As mentioned above, ompi/mpi/f77 and ompi/mpi/f90 no longer exist. All the fortran bindings implementations have been collated under ompi/mpi/fortran; each implementation has its own subdirectory:
{{{
ompi/mpi/fortran/
base/ - glue code
mpif-h/ - what used to be ompi/mpi/f77
use-mpi-tkr/ - what used to be ompi/mpi/f90
use-mpi-ignore-tkr/ - new mpi module implementation
use-mpi-f08/ - new mpi_f08 module implementation
}}}
There's also a prototype 6-function-MPI implementation under use-mpi-f08-desc that emulates the new F08 dimension(..) syntax that isn't fully available in Fortran compilers yet. We did that to prove it to ourselves that it could be done once the compilers fully support it. This directory/implementation will likely eventually replace the use-mpi-f08 version.
Other things that were done:
* ompi_info grew a few new output fields to describe what level of Fortran support is included
* Existing Fortran examples in examples/ were renamed; new mpi_f08 examples were added
* The old Fortran MPI libraries were renamed:
* libmpi_f77 -> libmpi_mpifh
* libmpi_f90 -> libmpi_usempi
* The configury for Fortran was consolidated and significantly slimmed down. Note that the F77 env variable is now IGNORED for configure; you should only use FC. Example:
{{{
shell$ ./configure CC=icc CXX=icpc FC=ifort ...
}}}
All of this work was done in a Mercurial branch off the SVN trunk, and hosted at Bitbucket. This branch has got to be one of OMPI's longest-running branches. Its first commit was Tue Apr 07 23:01:46 2009 -0400 -- it's over 3 years old! :-) We think we've pulled in all relevant changes from the OMPI trunk (e.g., Fortran implementations of the new MPI-3 MPROBE stuff for mpif.h, use mpi, and use mpi_f08, and the recent Fujitsu Fortran patches).
I anticipate some instability when we bring this stuff into the trunk, simply because it touches a LOT of code in the MPI layer in the OMPI code base. We'll try our best to make it as pain-free as possible, but please bear with us when it is committed.
This commit was SVN r26283.
2012-04-18 19:57:29 +04:00
|
|
|
#!/usr/bin/env perl
|
|
|
|
|
#
|
|
|
|
|
# Copyright (c) 2011 Cisco Systems, Inc. All rights reserved.
|
|
|
|
|
# $COPYRIGHT$
|
|
|
|
|
#
|
|
|
|
|
# Additional copyrights may follow
|
|
|
|
|
#
|
|
|
|
|
# $HEADER$
|
|
|
|
|
#
|
|
|
|
|
|
|
|
|
|
# This file generates common parameter values needed for the three
|
|
|
|
|
# MPI bindings. It creates the files:
|
|
|
|
|
#
|
|
|
|
|
# ompi/include/mpif-common.h
|
|
|
|
|
# ommp/mpi/fortran/use-mpi-f08/constants.h
|
|
|
|
|
|
|
|
|
|
use strict;
|
|
|
|
|
|
|
|
|
|
#----------------------------------------------------------------------------
|
|
|
|
|
|
|
|
|
|
# Read a file; return its contents in a string
|
|
|
|
|
|
|
|
|
|
sub read_file {
|
|
|
|
|
my ($filename_in) = @_;
|
|
|
|
|
|
|
|
|
|
open(FILE_IN, $filename_in) || die "Couldn't open file $filename_in";
|
|
|
|
|
my $fin;
|
|
|
|
|
$fin .= $_
|
|
|
|
|
while (<FILE_IN>);
|
|
|
|
|
close(FILE_IN);
|
|
|
|
|
|
|
|
|
|
return $fin;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#----------------------------------------------------------------------------
|
|
|
|
|
|
|
|
|
|
# Write an output file only if a) the output file does not already
|
|
|
|
|
# exist, or b) it exists, but its contents are different than $str.
|
|
|
|
|
|
|
|
|
|
sub write_file {
|
|
|
|
|
my ($filename_out, $str) = @_;
|
|
|
|
|
|
|
|
|
|
my $need_write = 0;
|
|
|
|
|
if (! -f $filename_out) {
|
|
|
|
|
$need_write = 1;
|
|
|
|
|
} else {
|
|
|
|
|
open(FILE_IN, $filename_out) || die "Couldn't open $filename_out";
|
|
|
|
|
my $tmp;
|
|
|
|
|
$tmp .= $_
|
|
|
|
|
while (<FILE_IN>);
|
|
|
|
|
close(FILE_IN);
|
|
|
|
|
if ($str ne $tmp) {
|
|
|
|
|
$need_write = 1;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if ($need_write) {
|
|
|
|
|
open(FILE_OUT, ">$filename_out") || die "Couldn't open $filename_out";
|
|
|
|
|
print FILE_OUT $str;
|
|
|
|
|
close(FILE_OUT);
|
|
|
|
|
print "created $filename_out\n";
|
|
|
|
|
} else {
|
|
|
|
|
print "$filename_out unchanged; not written\n";
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#----------------------------------------------------------------------------
|
|
|
|
|
|
|
|
|
|
print "creating Fortran header files (with common constants)...\n";
|
|
|
|
|
|
|
|
|
|
# Find the OMPI topdir. It is likely the pwd.
|
|
|
|
|
my $topdir;
|
|
|
|
|
if (-r "ompi/include/mpi.h.in") {
|
|
|
|
|
$topdir = ".";
|
|
|
|
|
} elsif (-r "include/mpi.h.in") {
|
|
|
|
|
$topdir = "..";
|
|
|
|
|
} elsif (-r "mpi.h.in") {
|
|
|
|
|
$topdir = "../..";
|
|
|
|
|
} else {
|
|
|
|
|
print "Please run this script from the Open MPI topdir or topdir/include/mpi\n";
|
|
|
|
|
print "Aborting.\n";
|
|
|
|
|
exit(1);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#----------------------------------------------------------------------------
|
|
|
|
|
|
|
|
|
|
my $MPI_COMM_WORLD = 0;
|
|
|
|
|
my $MPI_COMM_SELF = 1;
|
|
|
|
|
my $MPI_GROUP_EMPTY = 1;
|
|
|
|
|
my $MPI_ERRORS_ARE_FATAL = 1;
|
|
|
|
|
my $MPI_ERRORS_RETURN = 2;
|
|
|
|
|
|
2012-05-29 21:54:33 +04:00
|
|
|
#
|
|
|
|
|
# lookup table indices
|
|
|
|
|
#
|
|
|
|
|
my $MPI_MAX = 1;
|
|
|
|
|
my $MPI_MIN = 2;
|
|
|
|
|
my $MPI_SUM = 3;
|
|
|
|
|
my $MPI_PROD = 4;
|
|
|
|
|
my $MPI_LAND = 5;
|
|
|
|
|
my $MPI_BAND = 6;
|
|
|
|
|
my $MPI_LOR = 7;
|
|
|
|
|
my $MPI_BOR = 8;
|
|
|
|
|
my $MPI_LXOR = 9;
|
|
|
|
|
my $MPI_BXOR = 10;
|
|
|
|
|
my $MPI_MAXLOC = 11;
|
|
|
|
|
my $MPI_MINLOC = 12;
|
|
|
|
|
my $MPI_REPLACE = 13;
|
|
|
|
|
|
== Highlights ==
1. New mpifort wrapper compiler: you can utilize mpif.h, use mpi, and use mpi_f08 through this one wrapper compiler
1. mpif77 and mpif90 still exist, but are sym links to mpifort and may be removed in a future release
1. The mpi module has been re-implemented and is significantly "mo' bettah"
1. The mpi_f08 module offers many, many improvements over mpif.h and the mpi module
This stuff is coming from a VERY long-lived mercurial branch (3 years!); it'll almost certainly take a few SVN commits and a bunch of testing before I get it correctly committed to the SVN trunk.
== More details ==
Craig Rasmussen and I have been working with the MPI-3 Fortran WG and Fortran J3 committees for a long, long time to make a prototype MPI-3 Fortran bindings implementation. We think we're at a stable enough state to bring this stuff back to the trunk, with the goal of including it in OMPI v1.7.
Special thanks go out to everyone who has been incredibly patient and helpful to us in this journey:
* Rolf Rabenseifner/HLRS (mastermind/genius behind the entire MPI-3 Fortran effort)
* The Fortran J3 committee
* Tobias Burnus/gfortran
* Tony !Goetz/Absoft
* Terry !Donte/Oracle
* ...and probably others whom I'm forgetting :-(
There's still opportunities for optimization in the mpi_f08 implementation, but by and large, it is as far along as it can be until Fortran compilers start implementing the new F08 dimension(..) syntax.
Note that gfortran is currently unsupported for the mpi_f08 module and the new mpi module. gfortran users will a) fall back to the same mpi module implementation that is in OMPI v1.5.x, and b) not get the new mpi_f08 module. The gfortran maintainers are actively working hard to add the necessary features to support both the new mpi_f08 module and the new mpi module implementations. This will take some time.
As mentioned above, ompi/mpi/f77 and ompi/mpi/f90 no longer exist. All the fortran bindings implementations have been collated under ompi/mpi/fortran; each implementation has its own subdirectory:
{{{
ompi/mpi/fortran/
base/ - glue code
mpif-h/ - what used to be ompi/mpi/f77
use-mpi-tkr/ - what used to be ompi/mpi/f90
use-mpi-ignore-tkr/ - new mpi module implementation
use-mpi-f08/ - new mpi_f08 module implementation
}}}
There's also a prototype 6-function-MPI implementation under use-mpi-f08-desc that emulates the new F08 dimension(..) syntax that isn't fully available in Fortran compilers yet. We did that to prove it to ourselves that it could be done once the compilers fully support it. This directory/implementation will likely eventually replace the use-mpi-f08 version.
Other things that were done:
* ompi_info grew a few new output fields to describe what level of Fortran support is included
* Existing Fortran examples in examples/ were renamed; new mpi_f08 examples were added
* The old Fortran MPI libraries were renamed:
* libmpi_f77 -> libmpi_mpifh
* libmpi_f90 -> libmpi_usempi
* The configury for Fortran was consolidated and significantly slimmed down. Note that the F77 env variable is now IGNORED for configure; you should only use FC. Example:
{{{
shell$ ./configure CC=icc CXX=icpc FC=ifort ...
}}}
All of this work was done in a Mercurial branch off the SVN trunk, and hosted at Bitbucket. This branch has got to be one of OMPI's longest-running branches. Its first commit was Tue Apr 07 23:01:46 2009 -0400 -- it's over 3 years old! :-) We think we've pulled in all relevant changes from the OMPI trunk (e.g., Fortran implementations of the new MPI-3 MPROBE stuff for mpif.h, use mpi, and use mpi_f08, and the recent Fujitsu Fortran patches).
I anticipate some instability when we bring this stuff into the trunk, simply because it touches a LOT of code in the MPI layer in the OMPI code base. We'll try our best to make it as pain-free as possible, but please bear with us when it is committed.
This commit was SVN r26283.
2012-04-18 19:57:29 +04:00
|
|
|
#
|
|
|
|
|
# NULL "handles" (indices)
|
|
|
|
|
#
|
|
|
|
|
my $MPI_COMM_NULL = 2;
|
|
|
|
|
my $MPI_DATATYPE_NULL = 0;
|
|
|
|
|
my $MPI_ERRHANDLER_NULL = 0;
|
|
|
|
|
my $MPI_GROUP_NULL = 0;
|
|
|
|
|
my $MPI_INFO_NULL = 0;
|
|
|
|
|
my $MPI_MESSAGE_NULL = 0;
|
|
|
|
|
my $MPI_OP_NULL = 0;
|
|
|
|
|
my $MPI_REQUEST_NULL = 0;
|
|
|
|
|
my $MPI_WIN_NULL = 0;
|
|
|
|
|
my $MPI_MESSAGE_NULL = 0;
|
|
|
|
|
|
|
|
|
|
my $MPI_BYTE = 1;
|
|
|
|
|
my $MPI_PACKED = 2;
|
|
|
|
|
my $MPI_UB = 3;
|
|
|
|
|
my $MPI_LB = 4;
|
|
|
|
|
my $MPI_CHARACTER = 5;
|
|
|
|
|
my $MPI_LOGICAL = 6;
|
|
|
|
|
my $MPI_INTEGER = 7;
|
|
|
|
|
my $MPI_INTEGER1 = 8;
|
|
|
|
|
my $MPI_INTEGER2 = 9;
|
|
|
|
|
my $MPI_INTEGER4 = 10;
|
|
|
|
|
my $MPI_INTEGER8 = 11;
|
|
|
|
|
my $MPI_INTEGER16 = 12;
|
|
|
|
|
my $MPI_REAL = 13;
|
|
|
|
|
my $MPI_REAL4 = 14;
|
|
|
|
|
my $MPI_REAL8 = 15;
|
|
|
|
|
my $MPI_REAL16 = 16;
|
|
|
|
|
my $MPI_DOUBLE_PRECISION = 17;
|
|
|
|
|
my $MPI_COMPLEX = 18;
|
|
|
|
|
my $MPI_COMPLEX8 = 19;
|
|
|
|
|
my $MPI_COMPLEX16 = 20;
|
|
|
|
|
my $MPI_COMPLEX32 = 21;
|
|
|
|
|
my $MPI_DOUBLE_COMPLEX = 22;
|
|
|
|
|
my $MPI_2REAL = 23;
|
|
|
|
|
my $MPI_2DOUBLE_PRECISION = 24;
|
|
|
|
|
my $MPI_2INTEGER = 25;
|
|
|
|
|
my $MPI_2COMPLEX = 26;
|
|
|
|
|
my $MPI_2DOUBLE_COMPLEX = 27;
|
|
|
|
|
my $MPI_REAL2 = 28;
|
|
|
|
|
my $MPI_LOGICAL1 = 29;
|
|
|
|
|
my $MPI_LOGICAL2 = 30;
|
|
|
|
|
my $MPI_LOGICAL4 = 31;
|
|
|
|
|
my $MPI_LOGICAL8 = 32;
|
|
|
|
|
my $MPI_WCHAR = 33;
|
|
|
|
|
my $MPI_CHAR = 34;
|
|
|
|
|
my $MPI_UNSIGNED_CHAR = 35;
|
|
|
|
|
my $MPI_SIGNED_CHAR = 36;
|
|
|
|
|
my $MPI_SHORT = 37;
|
|
|
|
|
my $MPI_UNSIGNED_SHORT = 38;
|
|
|
|
|
my $MPI_INT = 39;
|
|
|
|
|
my $MPI_UNSIGNED = 40;
|
|
|
|
|
my $MPI_LONG = 41;
|
|
|
|
|
my $MPI_UNSIGNED_LONG = 42;
|
|
|
|
|
my $MPI_LONG_LONG_INT = 43;
|
|
|
|
|
my $MPI_UNSIGNED_LONG_LONG = 44;
|
|
|
|
|
my $MPI_FLOAT = 45;
|
|
|
|
|
my $MPI_DOUBLE = 46;
|
|
|
|
|
my $MPI_LONG_DOUBLE = 47;
|
|
|
|
|
my $MPI_FLOAT_INT = 48;
|
|
|
|
|
my $MPI_DOUBLE_INT = 49;
|
|
|
|
|
my $MPI_LONGDBL_INT = 50;
|
|
|
|
|
my $MPI_LONG_INT = 51;
|
|
|
|
|
my $MPI_2INT = 52;
|
|
|
|
|
my $MPI_SHORT_INT = 53;
|
|
|
|
|
my $MPI_CXX_BOOL = 54;
|
|
|
|
|
my $MPI_CXX_CPLEX = 55;
|
|
|
|
|
my $MPI_CXX_DBLCPLEX = 56;
|
|
|
|
|
my $MPI_CXX_LDBLCPLEX = 57;
|
|
|
|
|
my $MPI_INT8_T = 58;
|
|
|
|
|
my $MPI_UINT8_T = 59;
|
|
|
|
|
my $MPI_INT16_T = 60;
|
|
|
|
|
my $MPI_UINT16_T = 61;
|
|
|
|
|
my $MPI_INT32_T = 62;
|
|
|
|
|
my $MPI_UINT32_T = 63;
|
|
|
|
|
my $MPI_INT64_T = 64;
|
|
|
|
|
my $MPI_UINT64_T = 65;
|
|
|
|
|
my $MPI_AINT = 66;
|
|
|
|
|
my $MPI_OFFSET = 67;
|
|
|
|
|
|
|
|
|
|
#
|
|
|
|
|
# create ompi/include/mpif-common.h
|
|
|
|
|
#
|
|
|
|
|
|
|
|
|
|
my $filename_in = "$topdir/ompi/include/mpif-common.h.fin";
|
|
|
|
|
my $filename_out = $filename_in;
|
|
|
|
|
$filename_out =~ s/\.fin$//;
|
|
|
|
|
|
|
|
|
|
my $input = read_file($filename_in);
|
|
|
|
|
|
|
|
|
|
# Add warning to the top
|
|
|
|
|
$input = "! -*- fortran -*-
|
|
|
|
|
! WARNING! THIS IS A GENERATED FILE!!
|
|
|
|
|
! ANY EDITS YOU PUT HERE WILL BE LOST!
|
|
|
|
|
! ==> Instead, edit topdir/ompi/include/mpif-common.pl.
|
|
|
|
|
|
|
|
|
|
$input";
|
|
|
|
|
|
|
|
|
|
# Add more to the end
|
|
|
|
|
$input .= "!
|
|
|
|
|
! NULL 'handles' (indices)
|
|
|
|
|
!
|
|
|
|
|
|
|
|
|
|
integer MPI_GROUP_NULL, MPI_COMM_NULL, MPI_DATATYPE_NULL
|
|
|
|
|
integer MPI_REQUEST_NULL, MPI_OP_NULL, MPI_ERRHANDLER_NULL
|
|
|
|
|
integer MPI_INFO_NULL, MPI_WIN_NULL, MPI_MESSAGE_NULL
|
|
|
|
|
|
|
|
|
|
parameter (MPI_GROUP_NULL=$MPI_GROUP_NULL)
|
|
|
|
|
parameter (MPI_COMM_NULL=$MPI_COMM_NULL)
|
|
|
|
|
parameter (MPI_DATATYPE_NULL=$MPI_DATATYPE_NULL)
|
|
|
|
|
parameter (MPI_REQUEST_NULL=$MPI_REQUEST_NULL)
|
|
|
|
|
parameter (MPI_OP_NULL=$MPI_OP_NULL)
|
|
|
|
|
parameter (MPI_ERRHANDLER_NULL=$MPI_ERRHANDLER_NULL)
|
|
|
|
|
parameter (MPI_INFO_NULL=$MPI_INFO_NULL)
|
|
|
|
|
parameter (MPI_WIN_NULL=$MPI_WIN_NULL)
|
|
|
|
|
parameter (MPI_MESSAGE_NULL=$MPI_MESSAGE_NULL)
|
|
|
|
|
|
|
|
|
|
integer MPI_COMM_WORLD, MPI_COMM_SELF
|
|
|
|
|
integer MPI_GROUP_EMPTY
|
|
|
|
|
integer MPI_ERRORS_ARE_FATAL, MPI_ERRORS_RETURN
|
|
|
|
|
|
|
|
|
|
integer OMPI_MPI_COMM_WORLD
|
|
|
|
|
integer OMPI_MPI_COMM_SELF
|
|
|
|
|
|
|
|
|
|
parameter (MPI_COMM_WORLD=$MPI_COMM_WORLD)
|
|
|
|
|
parameter (MPI_COMM_SELF=$MPI_COMM_SELF)
|
|
|
|
|
parameter (MPI_GROUP_EMPTY=$MPI_GROUP_EMPTY)
|
|
|
|
|
parameter (MPI_ERRORS_ARE_FATAL=$MPI_ERRORS_ARE_FATAL)
|
|
|
|
|
parameter (MPI_ERRORS_RETURN=$MPI_ERRORS_RETURN)
|
|
|
|
|
|
2012-05-29 21:54:33 +04:00
|
|
|
!
|
|
|
|
|
! lookup table indices
|
|
|
|
|
!
|
|
|
|
|
|
|
|
|
|
integer MPI_MAX, MPI_MIN, MPI_SUM, MPI_PROD, MPI_LAND
|
|
|
|
|
integer MPI_BAND, MPI_LOR, MPI_BOR, MPI_LXOR, MPI_BXOR
|
|
|
|
|
integer MPI_MAXLOC, MPI_MINLOC, MPI_REPLACE
|
|
|
|
|
|
|
|
|
|
parameter (MPI_MAX=$MPI_MAX)
|
|
|
|
|
parameter (MPI_MIN=$MPI_MIN)
|
|
|
|
|
parameter (MPI_SUM=$MPI_SUM)
|
|
|
|
|
parameter (MPI_PROD=$MPI_PROD)
|
|
|
|
|
parameter (MPI_LAND=$MPI_LAND)
|
|
|
|
|
parameter (MPI_BAND=$MPI_BAND)
|
|
|
|
|
parameter (MPI_LOR=$MPI_LOR)
|
|
|
|
|
parameter (MPI_BOR=$MPI_BOR)
|
|
|
|
|
parameter (MPI_LXOR=$MPI_LXOR)
|
|
|
|
|
parameter (MPI_BXOR=$MPI_BXOR)
|
|
|
|
|
parameter (MPI_MAXLOC=$MPI_MAXLOC)
|
|
|
|
|
parameter (MPI_MINLOC=$MPI_MINLOC)
|
|
|
|
|
parameter (MPI_REPLACE=$MPI_REPLACE)
|
|
|
|
|
|
== Highlights ==
1. New mpifort wrapper compiler: you can utilize mpif.h, use mpi, and use mpi_f08 through this one wrapper compiler
1. mpif77 and mpif90 still exist, but are sym links to mpifort and may be removed in a future release
1. The mpi module has been re-implemented and is significantly "mo' bettah"
1. The mpi_f08 module offers many, many improvements over mpif.h and the mpi module
This stuff is coming from a VERY long-lived mercurial branch (3 years!); it'll almost certainly take a few SVN commits and a bunch of testing before I get it correctly committed to the SVN trunk.
== More details ==
Craig Rasmussen and I have been working with the MPI-3 Fortran WG and Fortran J3 committees for a long, long time to make a prototype MPI-3 Fortran bindings implementation. We think we're at a stable enough state to bring this stuff back to the trunk, with the goal of including it in OMPI v1.7.
Special thanks go out to everyone who has been incredibly patient and helpful to us in this journey:
* Rolf Rabenseifner/HLRS (mastermind/genius behind the entire MPI-3 Fortran effort)
* The Fortran J3 committee
* Tobias Burnus/gfortran
* Tony !Goetz/Absoft
* Terry !Donte/Oracle
* ...and probably others whom I'm forgetting :-(
There's still opportunities for optimization in the mpi_f08 implementation, but by and large, it is as far along as it can be until Fortran compilers start implementing the new F08 dimension(..) syntax.
Note that gfortran is currently unsupported for the mpi_f08 module and the new mpi module. gfortran users will a) fall back to the same mpi module implementation that is in OMPI v1.5.x, and b) not get the new mpi_f08 module. The gfortran maintainers are actively working hard to add the necessary features to support both the new mpi_f08 module and the new mpi module implementations. This will take some time.
As mentioned above, ompi/mpi/f77 and ompi/mpi/f90 no longer exist. All the fortran bindings implementations have been collated under ompi/mpi/fortran; each implementation has its own subdirectory:
{{{
ompi/mpi/fortran/
base/ - glue code
mpif-h/ - what used to be ompi/mpi/f77
use-mpi-tkr/ - what used to be ompi/mpi/f90
use-mpi-ignore-tkr/ - new mpi module implementation
use-mpi-f08/ - new mpi_f08 module implementation
}}}
There's also a prototype 6-function-MPI implementation under use-mpi-f08-desc that emulates the new F08 dimension(..) syntax that isn't fully available in Fortran compilers yet. We did that to prove it to ourselves that it could be done once the compilers fully support it. This directory/implementation will likely eventually replace the use-mpi-f08 version.
Other things that were done:
* ompi_info grew a few new output fields to describe what level of Fortran support is included
* Existing Fortran examples in examples/ were renamed; new mpi_f08 examples were added
* The old Fortran MPI libraries were renamed:
* libmpi_f77 -> libmpi_mpifh
* libmpi_f90 -> libmpi_usempi
* The configury for Fortran was consolidated and significantly slimmed down. Note that the F77 env variable is now IGNORED for configure; you should only use FC. Example:
{{{
shell$ ./configure CC=icc CXX=icpc FC=ifort ...
}}}
All of this work was done in a Mercurial branch off the SVN trunk, and hosted at Bitbucket. This branch has got to be one of OMPI's longest-running branches. Its first commit was Tue Apr 07 23:01:46 2009 -0400 -- it's over 3 years old! :-) We think we've pulled in all relevant changes from the OMPI trunk (e.g., Fortran implementations of the new MPI-3 MPROBE stuff for mpif.h, use mpi, and use mpi_f08, and the recent Fujitsu Fortran patches).
I anticipate some instability when we bring this stuff into the trunk, simply because it touches a LOT of code in the MPI layer in the OMPI code base. We'll try our best to make it as pain-free as possible, but please bear with us when it is committed.
This commit was SVN r26283.
2012-04-18 19:57:29 +04:00
|
|
|
integer MPI_BYTE, MPI_PACKED, MPI_UB, MPI_LB
|
|
|
|
|
integer MPI_CHARACTER, MPI_LOGICAL
|
|
|
|
|
integer MPI_INTEGER, MPI_INTEGER1, MPI_INTEGER2, MPI_INTEGER4
|
|
|
|
|
integer MPI_INTEGER8, MPI_INTEGER16
|
|
|
|
|
integer MPI_REAL, MPI_REAL2, MPI_REAL4, MPI_REAL8, MPI_REAL16
|
|
|
|
|
integer MPI_DOUBLE_PRECISION
|
|
|
|
|
integer MPI_COMPLEX, MPI_COMPLEX8, MPI_COMPLEX16, MPI_COMPLEX32
|
|
|
|
|
integer MPI_DOUBLE_COMPLEX
|
|
|
|
|
integer MPI_2REAL, MPI_2DOUBLE_PRECISION, MPI_2INTEGER
|
|
|
|
|
integer MPI_2COMPLEX, MPI_2DOUBLE_COMPLEX
|
|
|
|
|
|
|
|
|
|
! Note that MPI_LOGICALx are not defined by the MPI spec, but there are
|
|
|
|
|
! other MPI implementations that have them, so it's good for us to have
|
|
|
|
|
! as well.
|
|
|
|
|
|
|
|
|
|
integer MPI_LOGICAL1, MPI_LOGICAL2, MPI_LOGICAL4, MPI_LOGICAL8
|
|
|
|
|
|
|
|
|
|
! All other MPI types including the C and C++, as well as the
|
|
|
|
|
! ones defined in the MPI 2.2
|
|
|
|
|
|
|
|
|
|
integer MPI_WCHAR, MPI_CHAR
|
|
|
|
|
integer MPI_SIGNED_CHAR, MPI_UNSIGNED_CHAR
|
|
|
|
|
integer MPI_SHORT, MPI_UNSIGNED_SHORT
|
|
|
|
|
integer MPI_INT, MPI_UNSIGNED, MPI_LONG
|
|
|
|
|
integer MPI_UNSIGNED_LONG, MPI_LONG_LONG_INT
|
|
|
|
|
integer MPI_UNSIGNED_LONG_LONG
|
|
|
|
|
integer MPI_FLOAT, MPI_DOUBLE, MPI_LONG_DOUBLE
|
|
|
|
|
integer MPI_FLOAT_INT, MPI_DOUBLE_INT
|
|
|
|
|
integer MPI_LONGDBL_INT, MPI_LONG_INT
|
|
|
|
|
integer MPI_2INT, MPI_SHORT_INT
|
|
|
|
|
integer MPI_CXX_BOOL, MPI_CXX_CPLEX
|
|
|
|
|
integer MPI_CXX_DBLCPLEX, MPI_CXX_LDBLCPLEX
|
|
|
|
|
integer MPI_INT8_T, MPI_UINT8_T
|
|
|
|
|
integer MPI_INT16_T, MPI_UINT16_T
|
|
|
|
|
integer MPI_INT32_T, MPI_UINT32_T
|
|
|
|
|
integer MPI_INT64_T, MPI_UINT64_T
|
|
|
|
|
integer MPI_AINT, MPI_OFFSET
|
|
|
|
|
|
|
|
|
|
!
|
|
|
|
|
! Do NOT change the order of these parameters
|
|
|
|
|
!
|
|
|
|
|
|
|
|
|
|
parameter (MPI_BYTE = $MPI_BYTE)
|
|
|
|
|
parameter (MPI_PACKED = $MPI_PACKED)
|
|
|
|
|
parameter (MPI_UB = $MPI_UB)
|
|
|
|
|
parameter (MPI_LB = $MPI_LB)
|
|
|
|
|
parameter (MPI_CHARACTER = $MPI_CHARACTER)
|
|
|
|
|
parameter (MPI_LOGICAL = $MPI_LOGICAL)
|
|
|
|
|
parameter (MPI_INTEGER = $MPI_INTEGER)
|
|
|
|
|
parameter (MPI_INTEGER1 = $MPI_INTEGER1)
|
|
|
|
|
parameter (MPI_INTEGER2 = $MPI_INTEGER2)
|
|
|
|
|
parameter (MPI_INTEGER4 = $MPI_INTEGER4)
|
|
|
|
|
parameter (MPI_INTEGER8 = $MPI_INTEGER8)
|
|
|
|
|
parameter (MPI_INTEGER16 = $MPI_INTEGER16)
|
|
|
|
|
parameter (MPI_REAL = $MPI_REAL)
|
|
|
|
|
parameter (MPI_REAL4 = $MPI_REAL4)
|
|
|
|
|
parameter (MPI_REAL8 = $MPI_REAL8)
|
|
|
|
|
parameter (MPI_REAL16 = $MPI_REAL16)
|
|
|
|
|
parameter (MPI_DOUBLE_PRECISION = $MPI_DOUBLE_PRECISION)
|
|
|
|
|
parameter (MPI_COMPLEX = $MPI_COMPLEX)
|
|
|
|
|
parameter (MPI_COMPLEX8 = $MPI_COMPLEX8)
|
|
|
|
|
parameter (MPI_COMPLEX16 = $MPI_COMPLEX16)
|
|
|
|
|
parameter (MPI_COMPLEX32 = $MPI_COMPLEX32)
|
|
|
|
|
parameter (MPI_DOUBLE_COMPLEX = $MPI_DOUBLE_COMPLEX)
|
|
|
|
|
parameter (MPI_2REAL = $MPI_2REAL)
|
|
|
|
|
parameter (MPI_2DOUBLE_PRECISION = $MPI_2DOUBLE_PRECISION)
|
|
|
|
|
parameter (MPI_2INTEGER = $MPI_2INTEGER)
|
|
|
|
|
parameter (MPI_2COMPLEX = $MPI_2COMPLEX)
|
|
|
|
|
parameter (MPI_2DOUBLE_COMPLEX = $MPI_2DOUBLE_COMPLEX)
|
|
|
|
|
parameter (MPI_REAL2 = $MPI_REAL2)
|
|
|
|
|
parameter (MPI_LOGICAL1 = $MPI_LOGICAL1)
|
|
|
|
|
parameter (MPI_LOGICAL2 = $MPI_LOGICAL2)
|
|
|
|
|
parameter (MPI_LOGICAL4 = $MPI_LOGICAL4)
|
|
|
|
|
parameter (MPI_LOGICAL8 = $MPI_LOGICAL8)
|
|
|
|
|
parameter (MPI_WCHAR = $MPI_WCHAR)
|
|
|
|
|
parameter (MPI_CHAR = $MPI_CHAR)
|
|
|
|
|
parameter (MPI_UNSIGNED_CHAR = $MPI_UNSIGNED_CHAR)
|
|
|
|
|
parameter (MPI_SIGNED_CHAR = $MPI_SIGNED_CHAR)
|
|
|
|
|
parameter (MPI_SHORT = $MPI_SHORT)
|
|
|
|
|
parameter (MPI_UNSIGNED_SHORT = $MPI_UNSIGNED_SHORT)
|
|
|
|
|
parameter (MPI_INT = $MPI_INT)
|
|
|
|
|
parameter (MPI_UNSIGNED = $MPI_UNSIGNED)
|
|
|
|
|
parameter (MPI_LONG = $MPI_LONG)
|
|
|
|
|
parameter (MPI_UNSIGNED_LONG = $MPI_UNSIGNED_LONG)
|
|
|
|
|
parameter (MPI_LONG_LONG_INT = $MPI_LONG_LONG_INT)
|
|
|
|
|
parameter (MPI_UNSIGNED_LONG_LONG = $MPI_UNSIGNED_LONG_LONG)
|
|
|
|
|
parameter (MPI_FLOAT = $MPI_FLOAT)
|
|
|
|
|
parameter (MPI_DOUBLE = $MPI_DOUBLE)
|
|
|
|
|
parameter (MPI_LONG_DOUBLE = $MPI_LONG_DOUBLE)
|
|
|
|
|
parameter (MPI_FLOAT_INT = $MPI_FLOAT_INT)
|
|
|
|
|
parameter (MPI_DOUBLE_INT = $MPI_DOUBLE_INT)
|
|
|
|
|
parameter (MPI_LONGDBL_INT = $MPI_LONGDBL_INT)
|
|
|
|
|
parameter (MPI_LONG_INT = $MPI_LONG_INT)
|
|
|
|
|
parameter (MPI_2INT = $MPI_2INT)
|
|
|
|
|
parameter (MPI_SHORT_INT = $MPI_SHORT_INT)
|
|
|
|
|
parameter (MPI_CXX_BOOL = $MPI_CXX_BOOL)
|
|
|
|
|
parameter (MPI_CXX_CPLEX = $MPI_CXX_CPLEX)
|
|
|
|
|
parameter (MPI_CXX_DBLCPLEX = $MPI_CXX_DBLCPLEX)
|
|
|
|
|
parameter (MPI_CXX_LDBLCPLEX = $MPI_CXX_LDBLCPLEX)
|
|
|
|
|
parameter (MPI_INT8_T = $MPI_INT8_T)
|
|
|
|
|
parameter (MPI_UINT8_T = $MPI_UINT8_T)
|
|
|
|
|
parameter (MPI_INT16_T = $MPI_INT16_T)
|
|
|
|
|
parameter (MPI_UINT16_T = $MPI_UINT16_T)
|
|
|
|
|
parameter (MPI_INT32_T = $MPI_INT32_T)
|
|
|
|
|
parameter (MPI_UINT32_T = $MPI_UINT32_T)
|
|
|
|
|
parameter (MPI_INT64_T = $MPI_INT64_T)
|
|
|
|
|
parameter (MPI_UINT64_T = $MPI_UINT64_T)
|
|
|
|
|
parameter (MPI_AINT = $MPI_AINT)
|
|
|
|
|
parameter (MPI_OFFSET = $MPI_OFFSET)\n";
|
|
|
|
|
|
|
|
|
|
# Write the file
|
|
|
|
|
write_file($filename_out, $input);
|
|
|
|
|
|
|
|
|
|
#
|
|
|
|
|
# create ompi/mpi/fortran/use-mpi-f08/constants.h
|
|
|
|
|
#
|
|
|
|
|
|
|
|
|
|
$filename_in = "$topdir/ompi/mpi/fortran/use-mpi-f08/constants.h.fin";
|
|
|
|
|
$filename_out = $filename_in;
|
|
|
|
|
$filename_out =~ s/\.fin$//;
|
|
|
|
|
|
|
|
|
|
$input = read_file($filename_in);
|
|
|
|
|
|
|
|
|
|
# Add warning to the top
|
|
|
|
|
$input = "/* WARNING! THIS IS A GENERATED FILE!!
|
|
|
|
|
* ANY EDITS YOU PUT HERE WILL BE LOST!
|
|
|
|
|
* Instead, edit topdir/ompi/include/mpif-common.pl
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
$input";
|
|
|
|
|
|
|
|
|
|
# Add more to the end
|
|
|
|
|
$input .= "
|
|
|
|
|
#define OMPI_MPI_COMM_WORLD $MPI_COMM_WORLD
|
|
|
|
|
#define OMPI_MPI_COMM_SELF $MPI_COMM_SELF
|
|
|
|
|
#define OMPI_MPI_GROUP_EMPTY $MPI_GROUP_EMPTY
|
|
|
|
|
#define OMPI_MPI_ERRORS_ARE_FATAL $MPI_ERRORS_ARE_FATAL
|
|
|
|
|
#define OMPI_MPI_ERRORS_RETURN $MPI_ERRORS_RETURN
|
|
|
|
|
|
2012-05-29 21:54:33 +04:00
|
|
|
/*
|
|
|
|
|
* lookup table indices
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#define OMPI_MPI_MAX $MPI_MAX
|
|
|
|
|
#define OMPI_MPI_MIN $MPI_MIN
|
|
|
|
|
#define OMPI_MPI_SUM $MPI_SUM
|
|
|
|
|
#define OMPI_MPI_PROD $MPI_PROD
|
|
|
|
|
#define OMPI_MPI_LAND $MPI_LAND
|
|
|
|
|
#define OMPI_MPI_BAND $MPI_BAND
|
|
|
|
|
#define OMPI_MPI_LOR $MPI_LOR
|
|
|
|
|
#define OMPI_MPI_BOR $MPI_BOR
|
|
|
|
|
#define OMPI_MPI_LXOR $MPI_LXOR
|
|
|
|
|
#define OMPI_MPI_BXOR $MPI_BXOR
|
|
|
|
|
#define OMPI_MPI_MAXLOC $MPI_MAXLOC
|
|
|
|
|
#define OMPI_MPI_MINLOC $MPI_MINLOC
|
|
|
|
|
#define OMPI_MPI_REPLACE $MPI_REPLACE
|
|
|
|
|
|
== Highlights ==
1. New mpifort wrapper compiler: you can utilize mpif.h, use mpi, and use mpi_f08 through this one wrapper compiler
1. mpif77 and mpif90 still exist, but are sym links to mpifort and may be removed in a future release
1. The mpi module has been re-implemented and is significantly "mo' bettah"
1. The mpi_f08 module offers many, many improvements over mpif.h and the mpi module
This stuff is coming from a VERY long-lived mercurial branch (3 years!); it'll almost certainly take a few SVN commits and a bunch of testing before I get it correctly committed to the SVN trunk.
== More details ==
Craig Rasmussen and I have been working with the MPI-3 Fortran WG and Fortran J3 committees for a long, long time to make a prototype MPI-3 Fortran bindings implementation. We think we're at a stable enough state to bring this stuff back to the trunk, with the goal of including it in OMPI v1.7.
Special thanks go out to everyone who has been incredibly patient and helpful to us in this journey:
* Rolf Rabenseifner/HLRS (mastermind/genius behind the entire MPI-3 Fortran effort)
* The Fortran J3 committee
* Tobias Burnus/gfortran
* Tony !Goetz/Absoft
* Terry !Donte/Oracle
* ...and probably others whom I'm forgetting :-(
There's still opportunities for optimization in the mpi_f08 implementation, but by and large, it is as far along as it can be until Fortran compilers start implementing the new F08 dimension(..) syntax.
Note that gfortran is currently unsupported for the mpi_f08 module and the new mpi module. gfortran users will a) fall back to the same mpi module implementation that is in OMPI v1.5.x, and b) not get the new mpi_f08 module. The gfortran maintainers are actively working hard to add the necessary features to support both the new mpi_f08 module and the new mpi module implementations. This will take some time.
As mentioned above, ompi/mpi/f77 and ompi/mpi/f90 no longer exist. All the fortran bindings implementations have been collated under ompi/mpi/fortran; each implementation has its own subdirectory:
{{{
ompi/mpi/fortran/
base/ - glue code
mpif-h/ - what used to be ompi/mpi/f77
use-mpi-tkr/ - what used to be ompi/mpi/f90
use-mpi-ignore-tkr/ - new mpi module implementation
use-mpi-f08/ - new mpi_f08 module implementation
}}}
There's also a prototype 6-function-MPI implementation under use-mpi-f08-desc that emulates the new F08 dimension(..) syntax that isn't fully available in Fortran compilers yet. We did that to prove it to ourselves that it could be done once the compilers fully support it. This directory/implementation will likely eventually replace the use-mpi-f08 version.
Other things that were done:
* ompi_info grew a few new output fields to describe what level of Fortran support is included
* Existing Fortran examples in examples/ were renamed; new mpi_f08 examples were added
* The old Fortran MPI libraries were renamed:
* libmpi_f77 -> libmpi_mpifh
* libmpi_f90 -> libmpi_usempi
* The configury for Fortran was consolidated and significantly slimmed down. Note that the F77 env variable is now IGNORED for configure; you should only use FC. Example:
{{{
shell$ ./configure CC=icc CXX=icpc FC=ifort ...
}}}
All of this work was done in a Mercurial branch off the SVN trunk, and hosted at Bitbucket. This branch has got to be one of OMPI's longest-running branches. Its first commit was Tue Apr 07 23:01:46 2009 -0400 -- it's over 3 years old! :-) We think we've pulled in all relevant changes from the OMPI trunk (e.g., Fortran implementations of the new MPI-3 MPROBE stuff for mpif.h, use mpi, and use mpi_f08, and the recent Fujitsu Fortran patches).
I anticipate some instability when we bring this stuff into the trunk, simply because it touches a LOT of code in the MPI layer in the OMPI code base. We'll try our best to make it as pain-free as possible, but please bear with us when it is committed.
This commit was SVN r26283.
2012-04-18 19:57:29 +04:00
|
|
|
/*
|
|
|
|
|
* NULL 'handles' (indices)
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#define OMPI_MPI_GROUP_NULL $MPI_GROUP_NULL
|
|
|
|
|
#define OMPI_MPI_COMM_NULL $MPI_COMM_NULL
|
|
|
|
|
#define OMPI_MPI_DATATYPE_NULL $MPI_DATATYPE_NULL
|
|
|
|
|
#define OMPI_MPI_REQUEST_NULL $MPI_REQUEST_NULL
|
|
|
|
|
#define OMPI_MPI_OP_NULL $MPI_OP_NULL
|
|
|
|
|
#define OMPI_MPI_ERRHANDLER_NULL $MPI_ERRHANDLER_NULL
|
|
|
|
|
#define OMPI_MPI_INFO_NULL $MPI_INFO_NULL
|
|
|
|
|
#define OMPI_MPI_WIN_NULL $MPI_WIN_NULL
|
|
|
|
|
#define OMPI_MPI_MESSAGE_NULL $MPI_MESSAGE_NULL
|
|
|
|
|
|
|
|
|
|
#define OMPI_MPI_BYTE $MPI_BYTE
|
|
|
|
|
#define OMPI_MPI_PACKED $MPI_PACKED
|
|
|
|
|
#define OMPI_MPI_UB $MPI_UB
|
|
|
|
|
#define OMPI_MPI_LB $MPI_LB
|
|
|
|
|
#define OMPI_MPI_CHARACTER $MPI_CHARACTER
|
|
|
|
|
#define OMPI_MPI_LOGICAL $MPI_LOGICAL
|
|
|
|
|
#define OMPI_MPI_INTEGER $MPI_INTEGER
|
|
|
|
|
#define OMPI_MPI_INTEGER1 $MPI_INTEGER1
|
|
|
|
|
#define OMPI_MPI_INTEGER2 $MPI_INTEGER2
|
|
|
|
|
#define OMPI_MPI_INTEGER4 $MPI_INTEGER4
|
|
|
|
|
#define OMPI_MPI_INTEGER8 $MPI_INTEGER8
|
|
|
|
|
#define OMPI_MPI_INTEGER16 $MPI_INTEGER16
|
|
|
|
|
#define OMPI_MPI_REAL $MPI_REAL
|
|
|
|
|
#define OMPI_MPI_REAL4 $MPI_REAL4
|
|
|
|
|
#define OMPI_MPI_REAL8 $MPI_REAL8
|
|
|
|
|
#define OMPI_MPI_REAL16 $MPI_REAL16
|
|
|
|
|
#define OMPI_MPI_DOUBLE_PRECISION $MPI_DOUBLE_PRECISION
|
|
|
|
|
#define OMPI_MPI_COMPLEX $MPI_COMPLEX
|
|
|
|
|
#define OMPI_MPI_COMPLEX8 $MPI_COMPLEX8
|
|
|
|
|
#define OMPI_MPI_COMPLEX16 $MPI_COMPLEX16
|
|
|
|
|
#define OMPI_MPI_COMPLEX32 $MPI_COMPLEX32
|
|
|
|
|
#define OMPI_MPI_DOUBLE_COMPLEX $MPI_DOUBLE_COMPLEX
|
|
|
|
|
#define OMPI_MPI_2REAL $MPI_2REAL
|
|
|
|
|
#define OMPI_MPI_2DOUBLE_PRECISION $MPI_2DOUBLE_PRECISION
|
|
|
|
|
#define OMPI_MPI_2INTEGER $MPI_2INTEGER
|
|
|
|
|
#define OMPI_MPI_2COMPLEX $MPI_2COMPLEX
|
|
|
|
|
#define OMPI_MPI_2DOUBLE_COMPLEX $MPI_2DOUBLE_COMPLEX
|
|
|
|
|
#define OMPI_MPI_REAL2 $MPI_REAL2
|
|
|
|
|
#define OMPI_MPI_LOGICAL1 $MPI_LOGICAL1
|
|
|
|
|
#define OMPI_MPI_LOGICAL2 $MPI_LOGICAL2
|
|
|
|
|
#define OMPI_MPI_LOGICAL4 $MPI_LOGICAL4
|
|
|
|
|
#define OMPI_MPI_LOGICAL8 $MPI_LOGICAL8\n
|
|
|
|
|
|
|
|
|
|
#endif /* USE_MPI_F08_CONSTANTS_H */\n";
|
|
|
|
|
|
|
|
|
|
write_file($filename_out, $input);
|
|
|
|
|
exit(0);
|