George Bosilca 6a9514ee08 Make the code match the comment. I checked with Jelena, and based on the papers we ...

published this is the expected algorithm for the specified message and communicator
size.

This commit closes ticket #1330.

This commit was SVN r19563.

2008-09-15 23:28:40 +00:00

..

coll_tuned_allgather.c

- Make sure, that the *param_index are > 0 (here, we don't pass

2008-08-11 11:21:04 +00:00

coll_tuned_allgatherv.c

- Make sure, that the *param_index are > 0 (here, we don't pass

2008-08-11 11:21:04 +00:00

coll_tuned_allreduce.c

- Make sure, that the *param_index are > 0 (here, we don't pass

2008-08-11 11:21:04 +00:00

coll_tuned_alltoall.c

- Make sure, that the *param_index are > 0 (here, we don't pass

2008-08-11 11:21:04 +00:00

coll_tuned_alltoallv.c

- Make sure, that the *param_index are > 0 (here, we don't pass

2008-08-11 11:21:04 +00:00

coll_tuned_barrier.c

- Make sure, that the *param_index are > 0 (here, we don't pass

2008-08-11 11:21:04 +00:00

coll_tuned_bcast.c

- Make sure, that the *param_index are > 0 (here, we don't pass

2008-08-11 11:21:04 +00:00

coll_tuned_component.c

Fixes trac:1392, #1400

2008-07-28 22:40:57 +00:00

coll_tuned_decision_dynamic.c

Fixes trac:1392, #1400

2008-07-28 22:40:57 +00:00

coll_tuned_decision_fixed.c

Make the code match the comment. I checked with Jelena, and based on the papers we

2008-09-15 23:28:40 +00:00

coll_tuned_dynamic_file.c

Effectively revert the orte_output system and return to direct use of opal_output at all levels. Retain the orte_show_help subsystem to allow aggregation of show_help messages at the HNP.

2008-06-09 14:53:58 +00:00

coll_tuned_dynamic_file.h

Update to have public symbols be consistent; use new prefix rule

2005-12-22 13:49:33 +00:00

coll_tuned_dynamic_rules.c

Effectively revert the orte_output system and return to direct use of opal_output at all levels. Retain the orte_show_help subsystem to allow aggregation of show_help messages at the HNP.

2008-06-09 14:53:58 +00:00

coll_tuned_dynamic_rules.h

Adding variant of linear alltoall algorithm where the number of

2007-02-20 04:25:00 +00:00

coll_tuned_forced.c

Adding variant of linear alltoall algorithm where the number of

2007-02-20 04:25:00 +00:00

coll_tuned_forced.h

Adding variant of linear alltoall algorithm where the number of

2007-02-20 04:25:00 +00:00

coll_tuned_gather.c

- Make sure, that the *param_index are > 0 (here, we don't pass

2008-08-11 11:21:04 +00:00

coll_tuned_module.c

Fixes trac:1392, #1400

2008-07-28 22:40:57 +00:00

coll_tuned_reduce_scatter.c

- Make sure, that the *param_index are > 0 (here, we don't pass

2008-08-11 11:21:04 +00:00

coll_tuned_reduce.c

- Make sure, that the *param_index are > 0 (here, we don't pass

2008-08-11 11:21:04 +00:00

coll_tuned_scatter.c

- Make sure, that the *param_index are > 0 (here, we don't pass

2008-08-11 11:21:04 +00:00

coll_tuned_topo.c

Effectively revert the orte_output system and return to direct use of opal_output at all levels. Retain the orte_show_help subsystem to allow aggregation of show_help messages at the HNP.

2008-06-09 14:53:58 +00:00

coll_tuned_topo.h

added binomial tree based Gather alogrithm, passed IBM and Intel tests

2007-02-28 01:11:01 +00:00

coll_tuned_util.c

Effectively revert the orte_output system and return to direct use of opal_output at all levels. Retain the orte_show_help subsystem to allow aggregation of show_help messages at the HNP.

2008-06-09 14:53:58 +00:00

coll_tuned_util.h

Effectively revert the orte_output system and return to direct use of opal_output at all levels. Retain the orte_show_help subsystem to allow aggregation of show_help messages at the HNP.

2008-06-09 14:53:58 +00:00

coll_tuned.h

Fixes trac:1392, #1400

2008-07-28 22:40:57 +00:00

configure.params

Remove unneeded PARAM_INIT_FILE variable in configure.params files used by

2007-01-08 03:44:22 +00:00

Makefile.am

Add MPI_Alltoallv to tuned collectives and add a pairwise implementation of MPI_Alltoallv. However, do not change the default behavior for now. The only way to use new pairwise implementation is via mca parameters.

2008-05-07 02:31:24 +00:00