### Beowulf Cluster Help

Posted:

**Sun Mar 06, 2011 11:21 am**Hey, I'm working on a Beowulf Cluster for a class project, and we are using PelicanHPC for the OS ... anyhow the problem is, is we are trying to find a way to show the technologically impaired people that will be judging this project and currently we are trying to use srry i'm in a hurry

mpirun -np # octave -q --eval "kernal_example (2000, true)"

They say to set # at +1 of the number of cores you have ... however when i'm running/unplugging computers the number stays the same which doesn't make since ie: (these where all done with the same # of computers ( 3 computers 4 cores))

ser@pel1:~$ mpirun -np 3 octave -q --eval "kernel_example(2000,true)"

Just received fits for points 1 though 1000 out of 2000

Just received fits for points 1001 though 2000 out of 2000

Just received fits for points 1 though 1000 out of 2000

Just received fits for points 1001 though 2000 out of 2000

Just received fits for points 1 though 200 out of 400

Just received fits for points 201 though 400 out of 400

time for kernel regression example using 2000 data points on 3 nodes:

6.445394

time for kernel density example using 2000 data points on 3 nodes:

5.064756

time for bivariate kernel density example using 2000 data points on 3

nodes: 0.311769

==========================================================================

user@pel1:~$ mpirun -np 2 octave -q --eval "kernel_example(2000,true,

true)"

Just received fits for points 1 though 2000 out of 2000

Just received fits for points 1 though 2000 out of 2000

Just received fits for points 1 though 400 out of 400

time for kernel regression example using 2000 data points on 2 nodes:

7.768675

time for kernel density example using 2000 data points on 2 nodes:

7.539565

time for bivariate kernel density example using 2000 data points on 2

nodes: 0.519606

==========================================================================

user@pel1:~$ mpirun -np 9 octave -q --eval "kernel_example(2000,true,

true)"

Just received fits for points 1 though 2000 out of 2000

Just received fits for points 1 though 2000 out of 2000

Just received fits for points 1 though 400 out of 400

time for kernel regression example using 2000 data points on 2 nodes:

4.768675

time for kernel density example using 2000 data points on 2 nodes:

4.539565

time for bivariate kernel density example using 2000 data points on 2

nodes: 0.419606

mpirun -np # octave -q --eval "kernal_example (2000, true)"

They say to set # at +1 of the number of cores you have ... however when i'm running/unplugging computers the number stays the same which doesn't make since ie: (these where all done with the same # of computers ( 3 computers 4 cores))

ser@pel1:~$ mpirun -np 3 octave -q --eval "kernel_example(2000,true)"

Just received fits for points 1 though 1000 out of 2000

Just received fits for points 1001 though 2000 out of 2000

Just received fits for points 1 though 1000 out of 2000

Just received fits for points 1001 though 2000 out of 2000

Just received fits for points 1 though 200 out of 400

Just received fits for points 201 though 400 out of 400

time for kernel regression example using 2000 data points on 3 nodes:

6.445394

time for kernel density example using 2000 data points on 3 nodes:

5.064756

time for bivariate kernel density example using 2000 data points on 3

nodes: 0.311769

==========================================================================

user@pel1:~$ mpirun -np 2 octave -q --eval "kernel_example(2000,true,

true)"

Just received fits for points 1 though 2000 out of 2000

Just received fits for points 1 though 2000 out of 2000

Just received fits for points 1 though 400 out of 400

time for kernel regression example using 2000 data points on 2 nodes:

7.768675

time for kernel density example using 2000 data points on 2 nodes:

7.539565

time for bivariate kernel density example using 2000 data points on 2

nodes: 0.519606

==========================================================================

user@pel1:~$ mpirun -np 9 octave -q --eval "kernel_example(2000,true,

true)"

Just received fits for points 1 though 2000 out of 2000

Just received fits for points 1 though 2000 out of 2000

Just received fits for points 1 though 400 out of 400

time for kernel regression example using 2000 data points on 2 nodes:

4.768675

time for kernel density example using 2000 data points on 2 nodes:

4.539565

time for bivariate kernel density example using 2000 data points on 2

nodes: 0.419606