LAM/MPI General User's Mailing List Archives

From: Rohan Inamdar (inamroh_at_[hidden])
Date: 2003-05-12 21:21:59

• Next message: enrico.sirola_at_[hidden]: "Re: LAM: MPI_Comm_spawn problem - parent crashes if a child exit"
• Previous message: Pak, Anne O: "LAM: MPI_Barrier crashing MATLAB"
• Next in thread: Jeff Squyres: "Re: LAM: How to compute timings"
• Reply: Jeff Squyres: "Re: LAM: How to compute timings"

If my parent group contains only one process and I spawn a single process
using MPI_Comm_spawn(). How can i determine how much time did spawning take.

If call to MPI_Comm_spawn in above case returns, does it guarantee that
the process has been spawned and started on the destination machine. Documentation
only says that 'MPI_Comm_spawn' is collective over parent group.

How shall i determine how much time/overhead does spawning take over static
model of MPI 1.

Rohan Inamdar.

www.iit.edu/~inamroh

• Next message: enrico.sirola_at_[hidden]: "Re: LAM: MPI_Comm_spawn problem - parent crashes if a child exit"
• Previous message: Pak, Anne O: "LAM: MPI_Barrier crashing MATLAB"
• Next in thread: Jeff Squyres: "Re: LAM: How to compute timings"
• Reply: Jeff Squyres: "Re: LAM: How to compute timings"