Re: [O-MPI users] Thread-safety and daemon initiation

2005-02-15 Thread Jeff Squyres
Greetings Neil. In addition to what George and Mitch said: Open MPI will not require a full set of daemons like LAM/MPI does. You can have a full set if you wish, especially in environments where it makes sense (e.g., rsh/ssh, where there is little other infrastructure to do proper distribut

Re: [O-MPI users] Thread-safety and daemon initiation

2005-02-15 Thread Mitch Sukalski
Neil, Long time, no see! Greetings from rainy San Francisco to rainy Reading! Open MPI is designed to be thread-safe using pthread facilities with relatively fine grain access to critical regions. I haven't tried to mix Open MPI with OpenMP, but from what I understand of OpenMP, I believe it sh

Re: [O-MPI users] Thread-safety and daemon initiation

2005-02-15 Thread George Bosilca
Neil, Open-MPI support all thread models defined by the MPI standard if they are available on the target system. Few years ago I did some work with MPI and OpenMP and locking mechanism of existing MPI implementations was the performance killer. We have work hard to remove this bottleneck on O

[O-MPI users] Thread-safety and daemon initiation

2005-02-15 Thread Neil Storer
Hi, Two queries. 1) Is Open MPI thread-safe (assuming the system-supplied libraries are thread-safe)? In particular will it be possible to call MPI routines from OpenMP regions and will it be possible to use OpenMP directives within MPI tasks? 2) Does Open MPI use the same mechanism as LAM-