MPI_BOTTOM, Re: new version of misc chapter

Rolf Rabenseifner (Rabenseifner@RUS.Uni-Stuttgart.DE)
Tue, 27 Feb 1996 16:21:42 +0100 (MEZ)

To the proposals page 9 line 27 - page 10 line 5
I want to append an other perhaps waeker and better proposal
(it is a combination of proposal 1) and 2))


The virtual value of MPI_BOTTOM is used for MPI_ADDRESS.
This virtual value is the same for Fortran and C.
Therefore the results of MPI_ADDRESS are identical
in Fortran and C.
The real value of MPI_BOTTOM can by different in FORTRAN and C
and can be different to that virtual value.

This proposal can be implemented in the way of proposal one
(page 9 lines 28-38) or in the way of proposal two
(page 9 lines 39-43) and therefore the implementation
can choose between more programming effort in the
impelemtation (proposal one) or less efficiency (proposal two).

Therefore this is a proposal, that allows different values
but that can also avoid the additional test&branch.
The implementator can choose.

Please add this into the discussion list of Chap. 9.4.6.

Kind regards
Rolf


Rolf Rabenseifner (Computer Center )
Rechenzentrum Universitaet Stuttgart (University of Stuttgart)
Allmandring 30 Phone: ++49 711 6855530
D-70550 Stuttgart 80 FAX: ++49 711 6787626
Germany rabenseifner@rus.uni-stuttgart.de