openmpi

Brian Barrett a4fabb7e26 openib: Disable CUDA async by default

Disable async receive for CUDA under OpenIB.  While a performance
optimization, it also causes incorrect results for transfers
larger than the GPUDirect RDMA limit.  This change has been validated
and approved by Akshay.

References #3972

Signed-off-by: Brian Barrett <bbarrett@amazon.com>
(cherry picked from commit 9344afd485d23c41a18ab57d6df8550826c24b9e)
Signed-off-by: Brian Barrett <bbarrett@amazon.com>

2018-09-20 14:20:54 -07:00

class

opal/fifo: fix 128-bit atomic fifo on Power9

2018-07-10 15:37:11 -06:00

datatype

opal/dataype: add additional interface to retrieve more details about

2018-06-21 09:25:50 -05:00

dss

Update ORTE to support PMIx v3

2018-03-02 02:00:31 -08:00

etc

Correct the comment in the default MCA param template - we do not support a param called "component_path". The correct syntax is "mca_base_component_path"