Different datatypes with a single MPI broadcast, A program http://siber.cankaya.edu.tr/ozdogan/ParallelComputing/cfiles/code14.c code14.c that broadcast routine is used to communicate different datatypes with a single MPI broadcast (MPI_Bcast) call.
MPI datatypes are used.
All processes exit when a negative integer is read.
A SPMD program using broadcast and non-blocking receive. The http://siber.cankaya.edu.tr/ozdogan/ParallelComputing/cfiles/code16.c program consists of one sender process and up to 7 receiver processes.
The sender process broadcasts a message containing its identifier to all the other processes.
They receive the message and send an answer back, containing the hostname of the machine on which the process is running.
The receiving process waits for the first reply with MPI_Waitany, and accepts messages in the order they are received.
A SPMD program that uses MPI_Scatter. The http://siber.cankaya.edu.tr/ozdogan/ParallelComputing/cfiles/code17.c program should be run with an even number of processes.
Process zero initializes an array of integers x,
then distributes the array evenly among all processes using MPI_Scatter.
A SPMD program that uses MPI_Gather. The http://siber.cankaya.edu.tr/ozdogan/ParallelComputing/cfiles/code18.c program should be run with an even number of processes.
Each process initializes an array x of integers.
These arrays are collected to process zero using MPI_Gather and placed in an array y.
Timing comparison of processes and thread creation. Comparing timing results for the fork() subroutine and the pthreads_create() subroutine.
http://siber.cankaya.edu.tr/ozdogan/ParallelComputing/cfiles/code39.c code39.c,
http://siber.cankaya.edu.tr/ozdogan/ParallelComputing/cfiles/code40.c code40.c
Timings reflect 50,000 process/thread creations, were performed with the time utility (units are in seconds). Execute as
