Gilles Gouaillardet 60e91e890a coll/base: give a boost to ompi_coll_base_sendrecv_nonzero_actual() ... Based on current implementation it is faster to use a blocking send than the non-blocking version. Switch the exchange function used in the barrier to use the blocking version combined with the non-blocking version of the receive. This is similar to open-mpi/ompi@223d75595d		2016-08-04 13:31:07 +09:00
..
base	coll/base: give a boost to ompi_coll_base_sendrecv_nonzero_actual()	2016-08-04 13:31:07 +09:00
basic	coll/base: silence misc warning	2016-07-08 13:06:26 +09:00
cuda	Cleanup the memory handling for temporary buffers in	2015-12-02 20:42:18 -05:00
demo	Purge whitespace from the repo	2015-06-23 20:59:57 -07:00
fca	coll/fca: add missing #include <alloca.h>	2015-12-24 14:33:58 +09:00
hcoll	coll/hcoll mpi datatypes support	2016-07-29 10:06:39 +03:00
inter	coll/base: silence misc warning	2016-07-08 13:06:26 +09:00
libnbc	coll/libnbc: cleanup handling of the second temporary buffer in ireduce	2016-08-02 16:32:15 +09:00
ml	configury: clean the flex generated .c files	2016-06-01 11:13:31 +09:00
portals4	coll-portals4: check return value from reduce kary tree functions	2016-04-18 12:02:30 +00:00
self	Update the basic module to dynamically allocate the right	2015-10-08 12:00:41 -04:00
sm	Cleanup the memory handling for temporary buffers in	2015-12-02 20:42:18 -05:00
tuned	Fix comment.	2016-06-03 00:57:31 +09:00
coll.h	coll: do not cast way the const modifier when this is not necessary	2015-09-09 09:18:57 +09:00
Makefile.am	Purge whitespace from the repo	2015-06-23 20:59:57 -07:00