Mumps solver problems

This forum is for issues related to parallel processing
and OpenSees using the new interpreters OpenSeesSP and OpenSeesMP

Mumps solver problems

Postby dkkim » Mon Jul 07, 2008 7:57 am

Hellow..
I compiled the OS ver 1.7.5 with Mumps solver.
So, I tested it with a parallel example tcl script (ex4.tcl) like below..

# ---------------------------------------------------------------
set pid [getPID]
set np [getNP]
if {$np != 2} {
puts "ONLY WORKS FOR 2 PROCESSORS"
exit
}

model BasicBuilder -ndm 2 -ndf 2

uniaxialMaterial Elastic 1 3000

if {$pid == 0} {
node 1 0.0 0.0
node 4 72.0 96.0

fix 1 1 1

element truss 1 1 4 10.0 1

pattern Plain 1 "Linear" {
load 4 100 -50
}

} else {
node 2 144.0 0.0
node 3 168.0 0.0
node 4 72.0 96.0

fix 2 1 1
fix 3 1 1

element truss 2 2 4 5.0 1
element truss 3 3 4 5.0 1
}

# ------------------------------
# Create the recorders
# ------------------------------

recorder Node -file node4.out.$pid -time -node 4 -dof 1 2 disp

# ------------------------------
# Start of analysis generation
# ------------------------------

# Create the system of equation, a SPD using a band storage scheme
system Mumps
numberer ParallelPlain
constraints Plain
integrator LoadControl 0.1
test NormDispIncr 1.0e-12 3
algorithm Newton
analysis Static

analyze 10

print node 4

# ---------------------------------------------------------------


but there are some errors..

1.
the analysis was done at ONLY PID 1, not at PID 0.
so, ONE recorder(node4.out.1) file was created.
but, the result are zero.

and the Processor 0 goes on, does not stop.

so, I changed the Mumps solver to ParallelProfileSPD.
It created two result files(node4.out.0, node4.out.1) and reasonable values..

I think that it's Mumps problem.
What should I do to fix it?


2.
This is just a question about message.

what does this line mean?
Vector::Vector(int) - size 0 specified <= 0



3.
When I tested my large model, the mumps solver created many lines like below..

Vector::(loc) - loc 3554 outside range [0, 329
Vector::(loc) - loc 3555 outside range [0, 329
Vector::(loc) - loc 3556 outside range [0, 329

.... many lines.....

[3] MPI Abort by user Aborting program !
[3] Aborting program!


what is the problem...?


What I'd like to do is a large scale soil-structure interaction analysis

With UCSD nonlinear soil material , a parallel machine and Mumps solver.


Would you advise me about these things...?
dkkim
 
Posts: 9
Joined: Fri Sep 22, 2006 9:29 pm
Location: Seoul National University

Postby fmk » Mon Jul 07, 2008 2:09 pm

you need to update to version 2.0 .. the parallel
parts of 1.7.5 were still being modified and were not released for general consumption, so the version you have may contain bugs that have already been fixed.
fmk
Site Admin
 
Posts: 5860
Joined: Fri Jun 11, 2004 2:33 pm
Location: UC Berkeley


Return to Parallel Processing

Who is online

Users browsing this forum: No registered users and 1 guest