Hi,<br>
<br>
Thanks to Chan and Correa for their earlier replies that made me more insightful.<br>
<br>
I have compiled cpi.c (comes with MPICH2 example codes) using mpicc -o cpi.c cpi<br>
<br>
Ran it using two following commands - <br>
<br>
mpiexec ./cpi<br>
<br>
mpiexec -n 30 ./cpi<br>
<br>
In first case it says "Process 0 on localhost.localdomain" and also gives the value of PI and wall clock time (~ 10^-4 s).<br>
<br>
In second case it says "Process 0 from localhost.localdomain"<br>
"Process 1 on localhost.localdomain"<br>
...<br>
...<br>
"Process 5 on localhost.localdomain"<br>
<br>
etc. etc.<br>
<br>
(since, localhost.localdomain is the machine's default name)<br>
<br>
Since, my machine is a single box containing two processors (each having
6 cores); how do I know that it uses all the cores? Also the associated
calculation time is so small it is difficult to catch the processes
using system monitor.<br>
<br>
Is there any benchmarking program which will show how many cores were used for the calculation.<br>
<br>
Thanks in advance.<br>
<br>
Basak<br>