Towards an exhaustive test case list for MaPHyS

Table of Contents

This document describes several test cases for MaPHyS. The goal of it is to become an exhaustive test case file, including all possible tests in order to check the validity of a particular release of MaPHyS.

There are three kinds of test cases:

  1. unit tests for MaPHyS,
  2. simple real and complex matrices,
  3. and the deadly Alain Delon test cases.

To perform the tests with org-mode, we invite the reader to check the requirements section, describing the required installation of MaPHyS and containing a shell script for environment variables configuration and directories creation.

Once the installation and the directories are created, the entire document can be tangled with org-babel-tangle or <C-c C-v C-t>. This results in the creation of several input files for the Alain Delon test cases and also in the creation of a script named maphys-testcase.sh, allowing to launch all the test cases described in this document.

This last script can be run with org-mode by calling the sh code block in section Running the Non Regression Test Case script. It saves the results in another org-file, called results_testcase.org, also included in this document. It completes the section Running the Non Regression Test Case script with a brief report of the results of all the test cases.

This page has been generated from the following emacs org-mode file http://morse.gforge.inria.fr/maphys/maphys_test_cases.org and also available here:

svn checkout svn://scm.gforge.inria.fr/svnroot/morse/tutorials/maphys maphys_morse/ # or
svn checkout https://scm.gforge.inria.fr/anonscm/svn/morse/tutorials/maphys maphys_morse/

1 Useful links

2 Requirements

2.1 Installation

  • MaPHyS installation with both pastix and mumps
  • unit tests execs
  • examples execs
  • with mpich for MPI

To install MaPHyS, we highly recommend to use Spack. See http://morse.gforge.inria.fr/maphys/install-maphys-cluster.html for more information.

2.2 Configuration and directories creation

Fill variable MAPHYSBIN with the path to the examples of MaPHyS and execute the following sh code block:

#!/bin/bash
# MaPHyS non regression test cases script
# Author: HiePACS

# Set workdir for this file
export HOMERUN=${PWD}

# Set spack environment
#export SPACK_ROOT=path/to/your/spack/install
export SPACK_ROOT=${HOME}/lib/spack
. ${SPACK_ROOT}/share/spack/setup-env.sh
export MODULEPATH=$MODULEPATH:${SPACK_ROOT}/share/spack/modules/linux-x86_64

# Set path to maphys examples
export MAPHYSBIN=`spack location -i maphys`/bin

# Create directories
mkdir -p figures
mkdir -p testdir/unittest
mkdir -p testdir/matrices
mkdir -p testdir/alaindelon
mkdir -p testdir/alaindelon/ss
mkdir -p testdir/alaindelon/ms
mkdir -p testdir/alaindelon/sth
mkdir -p testdir/alaindelon/st
mkdir -p testdir/alaindelon/si
mkdir -p testdir/alaindelon/mt
mkdir -p testdir/alaindelon/mm
mkdir -p testdir/alaindelon/mi

Then, we need to load the Spack modules used to compile MaPHyS examples. Replace the module names if required.

# Load spack modules
#Hwloc
spack load hwloc

# MPI
spack load mpich

# Blas, lapack and scalapack
spack load eigen-blas
spack load netlib-lapack 
spack load netlib-scalapack

# Partitioners, scotch and metis
spack load scotch
spack load metis

#Direct sparse solvers
spack load pastix
spack load mumps

spack load maphys

3 Unit tests

To run MaPHyS unit tests, execute the following sh block.

# Begining of Unit tests
cd $HOMERUN

launch_ut () { 
mpirun -np $1 ${MAPHYSBIN}/$2mph_test_all > testdir/unittest/$2mph_ut_log_$1.out 2> testdir/unittest/$2mph_ut_log_$1.err &
wait
echo "#+BEGIN_EXAMPLE"
tail -n 5 testdir/unittest/$2mph_ut_log_$1.out
echo "See ${PWD}/testdir/unittest/$2mph_ut_log_$1.{out,err} for full details"
echo "#+END_EXAMPLE"
}

echo '** Unit tests results' 
echo "" 
echo "This section gives unit tests results for all arithmetics for several number of MPI processes."  


for i in 1 2 4; do
    echo "*** $i MPI processes"
    echo "**** Real simple precision"
    launch_ut $i 's'
    echo "**** Real double precision"
    launch_ut $i 'd'
    echo "**** Complex simple precision"
    launch_ut $i 'c' 
    echo "**** Complex double precision"
    launch_ut $i 'z'
done
rm proc00*

4 Real and complex matrices

To run MaPHyS matrix test for all arithmetics, execute the following sh block.

cd $HOMERUN

launch_mat () { 
if [ $1 == 1 ]; then
    ntail=16
else
    ntail=20
fi
timeout 20s mpirun -np $1 ${MAPHYSBIN}/$2mph_examplekv ${MAPHYSBIN}/$3 ${MAPHYSBIN} > testdir/matrices/$2mph_$1.out 2> testdir/matrices/$2mph_$1.err &
wait
echo "#+BEGIN_EXAMPLE"
tail -n $ntail testdir/matrices/$2mph_$1.out
echo "See ${PWD}/testdir/matrices/$2mph_$1.{out,err} for full details"
echo "#+END_EXAMPLE"
}

echo "** Beginning matrices test cases"
echo "*** With MUMPS"
for i in 1 2 4; do
    echo "**** $i MPI processes"
    echo "***** Real simple precision"
    launch_mat $i 's' 'real_bcsstk17_mumps.in'
    echo "***** Real double precision"
    launch_mat $i 'd' 'real_bcsstk17_mumps.in'
    echo "***** Complex simple precision"
    launch_mat $i 'c' 'complex_young1c_mumps.in'
    echo "***** Complex double precision"
    launch_mat $i 'z' 'complex_young1c_mumps.in'
done

echo "*** With PaStiX"
for i in 1 2 4; do
    echo "**** $i MPI processes"
    echo "***** Real simple precision"
    launch_mat $i 's' 'real_bcsstk17.in'
    echo "***** Real double precision"
    launch_mat $i 'd' 'real_bcsstk17.in'
    echo "***** Complex simple precision"
    launch_mat $i 'c' 'complex_young1c.in'
    echo "***** Complex double precision"
    launch_mat $i 'z' 'complex_young1c.in'
done

5 Alain Delon test cases

The Alain Delon test cases are designed to evaluate MaPHyS in the case of one or multiple subdomains possibly having themselves as neighbors, i.e. with both shared and non shared interface nodes. It also checks if MaPHyS is able to handle one or several domains with no interface points (see 5.2 test case for example).

5.1 Required matrix, rhs and input file

For each subdomains, we consider the same following matrices and rhs entries.

The input matrix is the following one:

%%MatrixMarket matrix coordinate real general
9 9 33
1 1 4.000000e+00
1 2 -1.000000e+00
1 4 -1.000000e+00
2 1 -1.000000e+00
2 2 4.000000e+00
2 3 -1.000000e+00
2 5 -1.000000e+00
3 2 -1.000000e+00
3 3 4.000000e+00
3 6 -1.000000e+00
4 1 -1.000000e+00
4 4 4.000000e+00
4 5 -1.000000e+00
4 7 -1.000000e+00
5 2 -1.000000e+00
5 4 -1.000000e+00
5 5 4.000000e+00
5 6 -1.000000e+00
5 8 -1.000000e+00
6 3 -1.000000e+00
6 5 -1.000000e+00
6 6 4.000000e+00
6 9 -1.000000e+00
7 4 -1.000000e+00
7 7 4.000000e+00
7 8 -1.000000e+00
8 5 -1.000000e+00
8 7 -1.000000e+00
8 8 4.000000e+00
8 9 -1.000000e+00
9 6 -1.000000e+00
9 8 -1.000000e+00
9 9 4.000000e+00

It's adjacency graph is given by Figure 1.

adglobal.png

Figure 1: Adjacency graph of all subdomains input matrix for the Alain Delon test cases

The right-hand-side is set such as the solution consist in a vector \((1\ 2\ \ldots \ 9)^T\). It is given by:

%%MatrixMarket matrix coordinate real general
9 1 9
1 1 -2.000000e+00
2 1 -1.000000e+00
3 1 4.000000e+00
4 1 3.000000e+00
5 1 0.
6 1 7.000000e+00
7 1 16.000000e+00
8 1 11.000000e+00
9 1 22.000000e+00

The input file is explicited in the following code bloc.

SYM = 0
ICNTL(4) = 3
ICNTL(5) = 1
ICNTL(6) = 1
ICNTL(7) = 4
ICNTL(13) = 1
ICNTL(20) = 1
ICNTL(21) = 1
ICNTL(24) = 500
ICNTL(26) = 500
ICNTL(27) = 1
ICNTL(22) = 3
RCNTL(21) = 1.0e-15
RCNTL(11) = 1.0e-6

5.2 Single Silent Alain Delon

The Single Silent Alain Delon test case consists in calling MaPHyS with a single domain and no interface nodes.

The single domain description is given by:

myndof=  9
mysizeintrf=  0
myinteface(:)= 

mynbvi= 0
myindexvi(:)= 
myptrindexvi(:)= 1 
myindexintrf(:)=

Figure 2 depicts the graph for this unique subdomain. Notice the interface is totally empty, as there is no red node on the figure.

adssglobal.png

Figure 2: Adjacency graph of the subdomain's input matrix for the Single Silent Alain Delon test case

To run the code, execute the following sh block.

# Launch single silent
echo "** Beginning Alain Delon test cases"
echo "*** Launching Alain Delon single silent"
mpirun -np 1 ${MAPHYSBIN}/dmph_examplerestart testdir/alaindelon/ss/param.in ${PWD}/testdir/alaindelon/ss > ${PWD}/testdir/alaindelon/ss/log.out &
wait
echo "#+BEGIN_EXAMPLE"
tail -n 15 ${PWD}/testdir/alaindelon/ss/log.out
echo "#+END_EXAMPLE"

5.3 Single Talkative Alain Delon

The Single Talkative Alain Delon test case consists in calling MaPHyS with one subdomain with the same matrix and rhs input than former test cases. But, in this case, the subdomain has a non empty interface.

Here, the interface nodes are associated to the unique subdomain, sharing these nodes with itself. We call these nodes self shared interface nodes. Hence, the subdomain is its own neighbor and it has to communicate with itself numerous times à la Alain Delon.

The matrix and rhs are duplicates of the single silent Alain Delon test case. So is the param file. Execute the following sh bloc to generate inputs.

# Single talkative sources
cp testdir/alaindelon/ss/maphys_local_matrix1.mtx    testdir/alaindelon/st/.
cp testdir/alaindelon/ss/maphys_local_rhs1.ijv       testdir/alaindelon/st/.
cp testdir/alaindelon/ss/param.in                    testdir/alaindelon/st/.

A tangle of the following code block gives the domain description:

myndof=  9
mysizeintrf=  3
myinteface(:)= 7 8 9

mynbvi= 1
myindexvi(:)= 0
myptrindexvi(:)= 1 4
myindexintrf(:)= 1 2 3

Figure 3 depicts the graph for this unique subdomain. Notice the interface is now non empty, as there are three red nodes (\(\{7,8,9\}\)) on the figure.

adstglobal.png

Figure 3: Adjacency graph of the subdomain's input matrix for the Single Talkative Alain Delon test case

To run the code, execute the following sh block.

# Launch single talkative
echo ""
echo "*** Launching Alain Delon single talkative"
mpirun -np 1 ${MAPHYSBIN}/dmph_examplerestart testdir/alaindelon/st/param.in ${PWD}/testdir/alaindelon/st > ${PWD}/testdir/alaindelon/st/log.out &
wait
echo "#+BEGIN_EXAMPLE"
tail -n 15 ${PWD}/testdir/alaindelon/st/log.out
echo "#+END_EXAMPLE"

5.4 Single Thinkative Alain Delon

The Single Silent Alain Delon test case consists in calling MaPHyS with a single domain and no interface nodes, but this time with interior points declared into the Schur. We name this nodes non shared interface nodes.

These nodes are going to be solved during the iterative part of MaPHyS solve phase. But, as they are not associated to any subdomain as interface nodes, they won't be associated to any neighbor, and hence won't be part of point to point communications.

However, these non shared interface nodes will be part of the scalar product computation during the iterative solve phase, because they are affected to the logic interface of the domain they belong by default.

The single domain description is given by:

myndof=  9
mysizeintrf=  2
myinteface(:)= 8 9

mynbvi= 0
myindexvi(:)= 
myptrindexvi(:)= 1 
myindexintrf(:)=

Figure 4 depicts the graph for this unique subdomain. Notice the interface is totally empty, as there is no red node on the figure. Notice also nodes \(\{8,9\}\) in green are the interior points that have been pushed into the Schur. These nodes will belong to the logic

adsthglobal.png

Figure 4: Adjacency graph of the subdomain's input matrix for the Single Silent Alain Delon test case

Execute the following sh bloc to generate inputs.

# Single thinkative sources
cp testdir/alaindelon/ss/maphys_local_matrix1.mtx    testdir/alaindelon/sth/.
cp testdir/alaindelon/ss/maphys_local_rhs1.ijv       testdir/alaindelon/sth/.
cp testdir/alaindelon/ss/param.in                    testdir/alaindelon/sth/.
# Launch single silent
echo "*** Launching Alain Delon single thinkative"
mpirun -np 1 ${MAPHYSBIN}/dmph_examplerestart testdir/alaindelon/sth/param.in ${PWD}/testdir/alaindelon/sth > ${PWD}/testdir/alaindelon/sth/log.out &
wait
echo "#+BEGIN_EXAMPLE"
tail -n 15 ${PWD}/testdir/alaindelon/sth/log.out
echo "#+END_EXAMPLE"

5.5 Single Inverse Alain Delon

To practice for his acting, Alain Delon sometimes speaks to himself in the mirror. The Inverse Alain Delon test case consists in calling MaPHyS with all nodes in the interface. It is also a way to test MaPHyS only with iterative solving.

The single domain description is given by:

myndof= 9
mysizeintrf= 9
myinteface(:)= 1 2 3 4 5 6 7 8 9

mynbvi= 1
myindexvi(:)= 0
myptrindexvi(:)= 1 10
myindexintrf(:)= 1 2 3 4 5 6 7 8 9

Execute the following sh bloc to generate inputs.

# Single inverse sources
cp testdir/alaindelon/ss/maphys_local_matrix1.mtx    testdir/alaindelon/si/.
cp testdir/alaindelon/ss/maphys_local_rhs1.ijv       testdir/alaindelon/si/.
cp testdir/alaindelon/ss/param.in                    testdir/alaindelon/si/.
# Launch single inverse
echo "*** Launching Alain Delon single inverse"
mpirun -np 1 ${MAPHYSBIN}/dmph_examplerestart testdir/alaindelon/si/param.in ${PWD}/testdir/alaindelon/si > ${PWD}/testdir/alaindelon/si/log.out &
wait
echo "#+BEGIN_EXAMPLE"
tail -n 15 ${PWD}/testdir/alaindelon/si/log.out
echo "#+END_EXAMPLE"

5.6 Multiple Silent Alain Delon

The Multiple Silent Alain Delon test case consists in calling MaPHyS with two and three subdomains with same matrices and rhs inputs and no interface nodes.

The matrices, rhs and domain descriptions are duplicates of the single silent Alain Delon test case. So is the param file.

Figure 5 depicts the global graph for the two subdomains with a global numbering of nodes. Notice the interface is totally empty, as there is no red node on the figure.

admsglobal.png

Figure 5: Adjacency graph of all subdomains input matrix for the Multiple Silent Alain Delon test case, global numbering

For the local orderings, see figure 6.

admsloc.png

Figure 6: Adjacency graph of all subdomains input matrix for the Multiple Silent Alain Delon test case, local numbering

Execute the following sh bloc to generate inputs.

# Multiple silent sources
cp testdir/alaindelon/ss/maphys_local_matrix1.mtx    testdir/alaindelon/ms/.
cp testdir/alaindelon/ss/maphys_local_matrix1.mtx    testdir/alaindelon/ms/maphys_local_matrix2.mtx
cp testdir/alaindelon/ss/maphys_local_matrix1.mtx    testdir/alaindelon/ms/maphys_local_matrix3.mtx
cp testdir/alaindelon/ss/maphys_local_rhs1.ijv       testdir/alaindelon/ms/.
cp testdir/alaindelon/ss/maphys_local_rhs1.ijv       testdir/alaindelon/ms/maphys_local_rhs2.ijv
cp testdir/alaindelon/ss/maphys_local_rhs1.ijv       testdir/alaindelon/ms/maphys_local_rhs3.ijv
cp testdir/alaindelon/ss/maphys_local_domain1.dom    testdir/alaindelon/ms/.
cp testdir/alaindelon/ss/maphys_local_domain1.dom    testdir/alaindelon/ms/maphys_local_domain2.dom
cp testdir/alaindelon/ss/maphys_local_domain1.dom    testdir/alaindelon/ms/maphys_local_domain3.dom
cp testdir/alaindelon/ss/param.in                    testdir/alaindelon/ms/.

To run the code, execute the following sh block.

# Launch multiple silent with 2 and 3 MPI processes
echo ""
echo "*** Launching Alain Delon multiple silent"
echo "**** 2 MPI processes"
mpirun -prepend-rank -np 2 ${MAPHYSBIN}/dmph_examplerestart testdir/alaindelon/ms/param.in ${PWD}/testdir/alaindelon/ms > ${PWD}/testdir/alaindelon/ms/log2.out &
wait
echo "#+BEGIN_EXAMPLE"
cat ${PWD}/testdir/alaindelon/ms/log2.out |grep '\[0\]' | tail -n 15
echo "#+END_EXAMPLE"
echo "**** 3 MPI processes"
mpirun -prepend-rank -np 3 ${MAPHYSBIN}/dmph_examplerestart testdir/alaindelon/ms/param.in ${PWD}/testdir/alaindelon/ms > ${PWD}/testdir/alaindelon/ms/log3.out &
wait
echo "#+BEGIN_EXAMPLE"
cat ${PWD}/testdir/alaindelon/ms/log3.out |grep '\[0\]'| tail -n 15
echo "#+END_EXAMPLE"

5.7 Multiple Talkative Alain Delon

The Multiple Talkative Alain Delon test case consists in calling MaPHyS with two subdomains with the same matrix and rhs input than former test cases.

For this test case, the two subdomains has a non empty interface, with both self shared and shared interface nodes. Hence, the subdomain are theirs own neighbors and have to communicate with themselves numerous times à la Alain Delon, and they also have to communicate with each other to handle the shared interface nodes.

The matrices and rhs are duplicates of the single silent Alain Delon test case. So is the param file. Execute the following sh bloc to generate inputs.

# Multiple talkative sources
cp testdir/alaindelon/ss/maphys_local_matrix1.mtx    testdir/alaindelon/mt/.
cp testdir/alaindelon/ss/maphys_local_matrix1.mtx    testdir/alaindelon/mt/maphys_local_matrix2.mtx
cp testdir/alaindelon/ss/maphys_local_rhs1.ijv       testdir/alaindelon/mt/.
cp testdir/alaindelon/ss/maphys_local_rhs1.ijv       testdir/alaindelon/mt/maphys_local_rhs2.ijv
cp testdir/alaindelon/ss/param.in                    testdir/alaindelon/mt/.

A tangle of the two following code blocs gives the domains description:

myndof=  9
mysizeintrf=  3
myinteface(:)= 7 8 9

mynbvi= 2
myindexvi(:)= 0 1
myptrindexvi(:)= 1 3 4
myindexintrf(:)= 1 2 3
myndof=  9
mysizeintrf=  2
myinteface(:)= 17 9

mynbvi= 2
myindexvi(:)= 1 0
myptrindexvi(:)= 1 2 3
myindexintrf(:)= 1 2

Figures 7 and 8 show respectively the global and local adjacency graph(s) for the two subdomains. Notice the interface here both shared with node 9 (global numbering, local one identical) in blue, and non shared with nodes \(\{7,8\}\) (global numbering \(\{7,8\}\)) for subdomain 1 and node \(\{17\}\) (global numbering, local one: \(\{8\}\)) for subdomain 2.

admtglobal.png

Figure 7: Adjacency graph of all subdomains input matrix for the Multiple Talkative Alain Delon test case, global numbering

For the local orderings, see figure 8.

admtloc.png

Figure 8: Adjacency graph of all subdomains input matrix for the Multiple Talkative Alain Delon test case, local numbering

To run the code, execute the following sh block.

# Launch multiple talkative
echo ""
echo "*** Launching Alain Delon multiple talkative"
mpirun -prepend-rank -np 2 ${MAPHYSBIN}/dmph_examplerestart testdir/alaindelon/mt/param.in ${PWD}/testdir/alaindelon/mt > ${PWD}/testdir/alaindelon/mt/log.out &
wait
echo "#+BEGIN_EXAMPLE"
cat ${PWD}/testdir/alaindelon/mt/log.out |grep '\[0\]' |tail -n 15
echo "#+END_EXAMPLE"

5.8 Multiple Mixed (silent and talkative) Alain Delon

The Multiple Mixed Alain Delon test case consists in calling MaPHyS with three subdomains with the same matrix, rhs and parameter input than former test cases. In that case, two subdomains has a non empty interface, with both shared and non shared interface nodes. The set of interface nodes is the same as for the 5.7 test case. See Figures 7 (global) and 8 (local) for the interface node list. The third one has an empty interface, with the same configuration as the 5.2 test case. See Figure 2 for its local adjacency graph.

Execute the following sh bloc to generate inputs.

# Multiple talkative sources
cp testdir/alaindelon/ss/maphys_local_matrix1.mtx    testdir/alaindelon/mm/.
cp testdir/alaindelon/ss/maphys_local_matrix1.mtx    testdir/alaindelon/mm/maphys_local_matrix2.mtx
cp testdir/alaindelon/ss/maphys_local_matrix1.mtx    testdir/alaindelon/mm/maphys_local_matrix3.mtx
cp testdir/alaindelon/ss/maphys_local_rhs1.ijv       testdir/alaindelon/mm/.
cp testdir/alaindelon/ss/maphys_local_rhs1.ijv       testdir/alaindelon/mm/maphys_local_rhs2.ijv
cp testdir/alaindelon/ss/maphys_local_rhs1.ijv       testdir/alaindelon/mm/maphys_local_rhs3.ijv
cp testdir/alaindelon/ss/param.in                    testdir/alaindelon/mm/.
cp testdir/alaindelon/mt/maphys_local_domain*        testdir/alaindelon/mm/.
cp testdir/alaindelon/ss/maphys_local_domain1.dom    testdir/alaindelon/mm/maphys_local_domain3.dom

To run the code, execute the following sh block.

# Launch multiple talkative
echo ""
echo "*** Launching Alain Delon multiple mixed"
mpirun -prepend-rank -np 3 ${MAPHYSBIN}/dmph_examplerestart testdir/alaindelon/mm/param.in ${PWD}/testdir/alaindelon/mm > ${PWD}/testdir/alaindelon/mm/log.out &
wait
echo "#+BEGIN_EXAMPLE"
cat ${PWD}/testdir/alaindelon/mm/log.out |grep '\[0\]' |tail -n 15
echo "#+END_EXAMPLE"

5.9 Multiple Inverse Alain Delon

The Multiple Inverse Alain Delon test case consists in calling MaPHyS with three subdomains as for the Multiple Mixed test case. This time all the processes have all their nodes in the interface: they are all Inverse Alain Delon.

Both processes 0 and 2 share two lines with each other and one line whith 1. Process 1 shares one line with 0, one line with 2, and one line with itself. Note that the global ordering is equivalent to the local ordering in this case.

The interface sharing is described in figure 9:

admiloc.png

Figure 9: Adjacency graph of all subdomains input matrix for the Multiple Inverse Alain Delon test case

Execute the following sh bloc to generate inputs.

# Multiple talkative sources
cp testdir/alaindelon/mm/maphys_local* testdir/alaindelon/mi/.
cp testdir/alaindelon/mm/param.in testdir/alaindelon/mi/.

A tangle of the three following code blocs gives the domains description:x

myndof=  9
mysizeintrf=  9
myinteface(:)= 1 2 3 4 5 6 7 8 9

mynbvi= 2
myindexvi(:)= 2 1
myptrindexvi(:)= 1 7 10
myindexintrf(:)= 1 2 3 4 5 6 7 8 9
myndof=  9
mysizeintrf=  9
myinteface(:)= 1 2 3 4 5 6 7 8 9

mynbvi= 3
myindexvi(:)= 0 1 2
myptrindexvi(:)= 1 4 7 10
myindexintrf(:)= 1 2 3 4 5 6 7 8 9
myndof=  9
mysizeintrf=  9
myinteface(:)= 1 2 3 4 5 6 7 8 9

mynbvi= 2
myindexvi(:)= 1 0
myptrindexvi(:)= 1 4 10
myindexintrf(:)= 1 2 3 4 5 6 7 8 9

To run the code, execute the following sh block.

# Launch multiple talkative
echo ""
echo "*** Launching Alain Delon multiple inverse"
mpirun -prepend-rank -np 3 ${MAPHYSBIN}/dmph_examplerestart testdir/alaindelon/mi/param.in ${PWD}/testdir/alaindelon/mi > ${PWD}/testdir/alaindelon/mi/log.out &
wait
echo "#+BEGIN_EXAMPLE"
cat ${PWD}/testdir/alaindelon/mi/log.out |grep '\[0\]' |tail -n 15
echo "#+END_EXAMPLE"

6 Running the test cases script

chmod 755 maphys-testcase.sh
./maphys-testcase.sh > results-testcases.org

6.1 Unit tests results

This section gives unit tests results for all arithmetics for several number of MPI processes.

6.1.1 1 MPI processes

  1. Real simple precision
    Summary of Launched tests
    PASSED =         29/        33
    FAILED =          0/        33
    OTHERS =          4/        33
    OTHERS = SKIPPED tests or those returning {K,X}{PASS,FAIL}
    See /home/mkuhn/morse/tutorials/maphys/testdir/unittest/smph_ut_log_1.{out,err} for full details
    
  2. Real double precision
    Summary of Launched tests
    PASSED =         29/        33
    FAILED =          0/        33
    OTHERS =          4/        33
    OTHERS = SKIPPED tests or those returning {K,X}{PASS,FAIL}
    See /home/mkuhn/morse/tutorials/maphys/testdir/unittest/dmph_ut_log_1.{out,err} for full details
    
  3. Complex simple precision
    Summary of Launched tests
    PASSED =         28/        33
    FAILED =          1/        33
    OTHERS =          4/        33
    OTHERS = SKIPPED tests or those returning {K,X}{PASS,FAIL}
    See /home/mkuhn/morse/tutorials/maphys/testdir/unittest/cmph_ut_log_1.{out,err} for full details
    
  4. Complex double precision
    Summary of Launched tests
    PASSED =         29/        33
    FAILED =          0/        33
    OTHERS =          4/        33
    OTHERS = SKIPPED tests or those returning {K,X}{PASS,FAIL}
    See /home/mkuhn/morse/tutorials/maphys/testdir/unittest/zmph_ut_log_1.{out,err} for full details
    

6.1.2 2 MPI processes

  1. Real simple precision
    Summary of Launched tests
    PASSED =         30/        33
    FAILED =          0/        33
    OTHERS =          3/        33
    OTHERS = SKIPPED tests or those returning {K,X}{PASS,FAIL}
    See /home/mkuhn/morse/tutorials/maphys/testdir/unittest/smph_ut_log_2.{out,err} for full details
    
  2. Real double precision
    Summary of Launched tests
    PASSED =         30/        33
    FAILED =          0/        33
    OTHERS =          3/        33
    OTHERS = SKIPPED tests or those returning {K,X}{PASS,FAIL}
    See /home/mkuhn/morse/tutorials/maphys/testdir/unittest/dmph_ut_log_2.{out,err} for full details
    
  3. Complex simple precision
    Summary of Launched tests
    PASSED =         29/        33
    FAILED =          1/        33
    OTHERS =          3/        33
    OTHERS = SKIPPED tests or those returning {K,X}{PASS,FAIL}
    See /home/mkuhn/morse/tutorials/maphys/testdir/unittest/cmph_ut_log_2.{out,err} for full details
    
  4. Complex double precision
    Summary of Launched tests
    PASSED =         30/        33
    FAILED =          0/        33
    OTHERS =          3/        33
    OTHERS = SKIPPED tests or those returning {K,X}{PASS,FAIL}
    See /home/mkuhn/morse/tutorials/maphys/testdir/unittest/zmph_ut_log_2.{out,err} for full details
    

6.1.3 4 MPI processes

  1. Real simple precision
    Summary of Launched tests
    PASSED =         30/        33
    FAILED =          0/        33
    OTHERS =          3/        33
    OTHERS = SKIPPED tests or those returning {K,X}{PASS,FAIL}
    See /home/mkuhn/morse/tutorials/maphys/testdir/unittest/smph_ut_log_4.{out,err} for full details
    
  2. Real double precision
    Summary of Launched tests
    PASSED =         30/        33
    FAILED =          0/        33
    OTHERS =          3/        33
    OTHERS = SKIPPED tests or those returning {K,X}{PASS,FAIL}
    See /home/mkuhn/morse/tutorials/maphys/testdir/unittest/dmph_ut_log_4.{out,err} for full details
    
  3. Complex simple precision
    Summary of Launched tests
    PASSED =         29/        33
    FAILED =          1/        33
    OTHERS =          3/        33
    OTHERS = SKIPPED tests or those returning {K,X}{PASS,FAIL}
    See /home/mkuhn/morse/tutorials/maphys/testdir/unittest/cmph_ut_log_4.{out,err} for full details
    
  4. Complex double precision
    Summary of Launched tests
    PASSED =         30/        33
    FAILED =          0/        33
    OTHERS =          3/        33
    OTHERS = SKIPPED tests or those returning {K,X}{PASS,FAIL}
    See /home/mkuhn/morse/tutorials/maphys/testdir/unittest/zmph_ut_log_4.{out,err} for full details
    

6.2 Beginning matrices test cases

6.2.1 With MUMPS

  1. 1 MPI processes
    1. Real simple precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     5.1121389328E-07
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     9.9527072906E-01 |     1.0000000000E+00
      |sol( 3)| |     9.9960899353E-01 |     1.0000000000E+00
      |sol( 4)| |     9.9998998642E-01 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     9.9999624491E-01 |     1.0000000000E+00
      |sol( 9)| |     9.9921232462E-01 |     1.0000000000E+00
      |sol(10)| |     9.9994981289E-01 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/smph_1.{out,err} for full details
      
    2. Real double precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     8.5094134297E-16
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 3)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 4)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000000000E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/dmph_1.{out,err} for full details
      
    3. Complex simple precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     1.5715167098E-06
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     9.9999731779E-01 |     1.0000000000E+00
      |sol( 2)| |     9.9999529123E-01 |     1.0000000000E+00
      |sol( 3)| |     9.9999433756E-01 |     1.0000000000E+00
      |sol( 4)| |     9.9999433756E-01 |     1.0000000000E+00
      |sol( 5)| |     9.9999392033E-01 |     1.0000000000E+00
      |sol( 6)| |     9.9999707937E-01 |     1.0000000000E+00
      |sol( 7)| |     9.9999922514E-01 |     1.0000000000E+00
      |sol( 8)| |     1.0000002384E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000007153E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000001192E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/cmph_1.{out,err} for full details
      
    4. Complex double precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     2.8772460612E-15
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 3)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 4)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000000000E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/zmph_1.{out,err} for full details
      
  2. 2 MPI processes
    1. Real simple precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     4.6878638969E-07
      Backward error centralized:               |A.x-b|/|b|    norm2 =     4.3744679562E-07
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     7.7885488422E-04
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     6.4104795456E-03
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     9.9358952045E-01 |     1.0000000000E+00
      |sol( 3)| |     9.9964624643E-01 |     1.0000000000E+00
      |sol( 4)| |     9.9998331070E-01 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     9.9999874830E-01 |     1.0000000000E+00
      |sol( 9)| |     9.9910581112E-01 |     1.0000000000E+00
      |sol(10)| |     9.9994987249E-01 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/smph_2.{out,err} for full details
      
    2. Real double precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     8.7706170565E-16
      Backward error centralized:               |A.x-b|/|b|    norm2 =     8.6252486209E-16
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     7.9101354762E-13
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     3.9599434842E-12
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 3)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 4)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000000000E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/dmph_2.{out,err} for full details
      
    3. Complex simple precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     1.6153591369E-06
      Backward error centralized:               |A.x-b|/|b|    norm2 =     1.6164008871E-06
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     3.4466041396E-06
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     1.4285686120E-05
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000001192E+00 |     1.0000000000E+00
      |sol( 2)| |     1.0000001192E+00 |     1.0000000000E+00
      |sol( 3)| |     1.0000001192E+00 |     1.0000000000E+00
      |sol( 4)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 5)| |     9.9999964237E-01 |     1.0000000000E+00
      |sol( 6)| |     9.9999940395E-01 |     1.0000000000E+00
      |sol( 7)| |     9.9999922514E-01 |     1.0000000000E+00
      |sol( 8)| |     9.9999946356E-01 |     1.0000000000E+00
      |sol( 9)| |     9.9999970198E-01 |     1.0000000000E+00
      |sol(10)| |     1.0000002384E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/cmph_2.{out,err} for full details
      
    4. Complex double precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     3.6382688138E-15
      Backward error centralized:               |A.x-b|/|b|    norm2 =     3.6371517752E-15
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     7.0206294616E-15
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     2.5408878587E-14
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 3)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 4)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000000000E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/zmph_2.{out,err} for full details
      
  3. 4 MPI processes
    1. Real simple precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     4.2773375053E-07
      Backward error centralized:               |A.x-b|/|b|    norm2 =     4.3569467232E-07
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     1.8991274039E-04
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     1.1859536171E-03
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     9.9881404638E-01 |     1.0000000000E+00
      |sol( 3)| |     9.9971193075E-01 |     1.0000000000E+00
      |sol( 4)| |     9.9999719858E-01 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000014305E+00 |     1.0000000000E+00
      |sol( 9)| |     9.9947327375E-01 |     1.0000000000E+00
      |sol(10)| |     9.9997735023E-01 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/smph_4.{out,err} for full details
      
    2. Real double precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     1.1256580765E-03
      Backward error centralized:               |A.x-b|/|b|    norm2 =     1.1256580765E-03
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     4.4715593293E-01
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     3.8102253126E+00
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     1.9521673264E+00 |     1.0000000000E+00
      |sol( 3)| |     2.7343756658E-01 |     1.0000000000E+00
      |sol( 4)| |     9.9414817579E-01 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     9.9922751503E-01 |     1.0000000000E+00
      |sol( 9)| |     3.7709798865E-01 |     1.0000000000E+00
      |sol(10)| |     9.8734065485E-01 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/dmph_4.{out,err} for full details
      
    3. Complex simple precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     8.8334554515E-06
      Backward error centralized:               |A.x-b|/|b|    norm2 =     8.8389509782E-06
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     3.1059452569E-05
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     1.1788320990E-04
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     9.9999982119E-01 |     1.0000000000E+00
      |sol( 2)| |     9.9999958277E-01 |     1.0000000000E+00
      |sol( 3)| |     9.9999934435E-01 |     1.0000000000E+00
      |sol( 4)| |     9.9999928474E-01 |     1.0000000000E+00
      |sol( 5)| |     9.9999904633E-01 |     1.0000000000E+00
      |sol( 6)| |     9.9999910593E-01 |     1.0000000000E+00
      |sol( 7)| |     9.9999940395E-01 |     1.0000000000E+00
      |sol( 8)| |     1.0000001192E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000008345E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000015497E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/cmph_4.{out,err} for full details
      
    4. Complex double precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     1.5243262505E-06
      Backward error centralized:               |A.x-b|/|b|    norm2 =     1.5243262505E-06
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     5.7072286016E-06
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     1.8157479909E-05
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     9.9999955107E-01 |     1.0000000000E+00
      |sol( 2)| |     9.9999924641E-01 |     1.0000000000E+00
      |sol( 3)| |     9.9999917948E-01 |     1.0000000000E+00
      |sol( 4)| |     9.9999936127E-01 |     1.0000000000E+00
      |sol( 5)| |     9.9999971880E-01 |     1.0000000000E+00
      |sol( 6)| |     1.0000001220E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000004260E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000005110E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000003050E+00 |     1.0000000000E+00
      |sol(10)| |     9.9999978532E-01 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/zmph_4.{out,err} for full details
      

6.2.2 With PaStiX

  1. 1 MPI processes
    1. Real simple precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     4.8343901525E-07
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     1.0029803514E+00 |     1.0000000000E+00
      |sol( 3)| |     1.0001705885E+00 |     1.0000000000E+00
      |sol( 4)| |     1.0000087023E+00 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000011921E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0001369715E+00 |     1.0000000000E+00
      |sol(10)| |     9.9999272823E-01 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/smph_1.{out,err} for full details
      
    2. Real double precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     2.4939212503E-15
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 3)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 4)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000000000E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/dmph_1.{out,err} for full details
      
    3. Complex simple precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     1.8139428282E-06
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     9.9999850988E-01 |     1.0000000000E+00
      |sol( 2)| |     9.9999868870E-01 |     1.0000000000E+00
      |sol( 3)| |     9.9999880791E-01 |     1.0000000000E+00
      |sol( 4)| |     9.9999856949E-01 |     1.0000000000E+00
      |sol( 5)| |     9.9999827147E-01 |     1.0000000000E+00
      |sol( 6)| |     9.9999791384E-01 |     1.0000000000E+00
      |sol( 7)| |     9.9999791384E-01 |     1.0000000000E+00
      |sol( 8)| |     9.9999809265E-01 |     1.0000000000E+00
      |sol( 9)| |     9.9999815226E-01 |     1.0000000000E+00
      |sol(10)| |     9.9999839067E-01 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/cmph_1.{out,err} for full details
      
    4. Complex double precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     8.9229720501E-15
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 3)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 4)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000000000E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/zmph_1.{out,err} for full details
      
  2. 2 MPI processes
    1. Real simple precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     4.5791992166E-07
      Backward error centralized:               |A.x-b|/|b|    norm2 =     4.8302257717E-07
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     1.1503852442E-03
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     9.6840858459E-03
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     9.9031591415E-01 |     1.0000000000E+00
      |sol( 3)| |     9.9985152483E-01 |     1.0000000000E+00
      |sol( 4)| |     9.9997776747E-01 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     9.9999475479E-01 |     1.0000000000E+00
      |sol( 9)| |     9.9922978878E-01 |     1.0000000000E+00
      |sol(10)| |     9.9995970726E-01 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/smph_2.{out,err} for full details
      
    2. Real double precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     1.0238770245E-15
      Backward error centralized:               |A.x-b|/|b|    norm2 =     9.9851427176E-16
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     1.5140121371E-12
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     9.9409369625E-12
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 3)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 4)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000000000E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/dmph_2.{out,err} for full details
      
    3. Complex simple precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     3.9392433479E-06
      Backward error centralized:               |A.x-b|/|b|    norm2 =     3.9393373409E-06
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     1.3381097808E-05
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     5.0829792599E-05
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000003576E+00 |     1.0000000000E+00
      |sol( 2)| |     1.0000004768E+00 |     1.0000000000E+00
      |sol( 3)| |     1.0000003576E+00 |     1.0000000000E+00
      |sol( 4)| |     1.0000001192E+00 |     1.0000000000E+00
      |sol( 5)| |     9.9999976158E-01 |     1.0000000000E+00
      |sol( 6)| |     9.9999934435E-01 |     1.0000000000E+00
      |sol( 7)| |     9.9999922514E-01 |     1.0000000000E+00
      |sol( 8)| |     9.9999958277E-01 |     1.0000000000E+00
      |sol( 9)| |     1.0000002384E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000009537E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/cmph_2.{out,err} for full details
      
    4. Complex double precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     6.2785446538E-15
      Backward error centralized:               |A.x-b|/|b|    norm2 =     6.2700068431E-15
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     1.8866271818E-14
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     9.4034440745E-14
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 3)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 4)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000000000E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/zmph_2.{out,err} for full details
      
  3. 4 MPI processes
    1. Real simple precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     4.3804291101E-07
      Backward error centralized:               |A.x-b|/|b|    norm2 =     4.1823925682E-07
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     1.3416178562E-03
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     1.0967075825E-02
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     9.8903292418E-01 |     1.0000000000E+00
      |sol( 3)| |     9.9975860119E-01 |     1.0000000000E+00
      |sol( 4)| |     9.9997645617E-01 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     9.9999231100E-01 |     1.0000000000E+00
      |sol( 9)| |     9.9884200096E-01 |     1.0000000000E+00
      |sol(10)| |     9.9994122982E-01 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/smph_4.{out,err} for full details
      
    2. Real double precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     8.7514191449E-16
      Backward error centralized:               |A.x-b|/|b|    norm2 =     8.5557076461E-16
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     1.8638980214E-12
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     1.3963497025E-11
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 2)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 3)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 4)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 5)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 6)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000000000E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000000000E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/dmph_4.{out,err} for full details
      
    3. Complex simple precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     2.0685329061E-06
      Backward error centralized:               |A.x-b|/|b|    norm2 =     2.0688524281E-06
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     5.9716841841E-06
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     2.8661777833E-05
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     9.9999880791E-01 |     1.0000000000E+00
      |sol( 2)| |     9.9999815226E-01 |     1.0000000000E+00
      |sol( 3)| |     9.9999803305E-01 |     1.0000000000E+00
      |sol( 4)| |     9.9999862909E-01 |     1.0000000000E+00
      |sol( 5)| |     9.9999964237E-01 |     1.0000000000E+00
      |sol( 6)| |     1.0000005960E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000014305E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000015497E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000007153E+00 |     1.0000000000E+00
      |sol(10)| |     1.0000005960E+00 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/cmph_4.{out,err} for full details
      
    4. Complex double precision
       == ESTIMATION OF COMMITED ERROR ==
      Backward error:                           |A.x-b|/|b|    norm2 =     1.5243262503E-06
      Backward error centralized:               |A.x-b|/|b|    norm2 =     1.5243262503E-06
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*|    norm2 =     5.7072285849E-06
      Computed (x) vs analytical (x*) solution: |x-x*|/|x*| norm_inf =     1.8157479864E-05
       
       == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
                |         Computed (x) |     Theoretical (x*)
      _____________________________________________________
      |sol( 1)| |     9.9999955107E-01 |     1.0000000000E+00
      |sol( 2)| |     9.9999924641E-01 |     1.0000000000E+00
      |sol( 3)| |     9.9999917948E-01 |     1.0000000000E+00
      |sol( 4)| |     9.9999936127E-01 |     1.0000000000E+00
      |sol( 5)| |     9.9999971880E-01 |     1.0000000000E+00
      |sol( 6)| |     1.0000001220E+00 |     1.0000000000E+00
      |sol( 7)| |     1.0000004260E+00 |     1.0000000000E+00
      |sol( 8)| |     1.0000005110E+00 |     1.0000000000E+00
      |sol( 9)| |     1.0000003050E+00 |     1.0000000000E+00
      |sol(10)| |     9.9999978532E-01 |     1.0000000000E+00
      See /home/mkuhn/morse/tutorials/maphys/testdir/matrices/zmph_4.{out,err} for full details
      

6.3 Beginning Alain Delon test cases

6.3.1 Launching Alain Delon single silent

 == ESTIMATION OF COMMITED ERROR ==
 Backward error: |A.x-b|/|b| =    2.0990169310970903E-016
 
 == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
          |         Computed (x) | 
_____________________________________________________
|sol( 1)| |     1.0000000000E+00 | 
|sol( 2)| |     2.0000000000E+00 | 
|sol( 3)| |     3.0000000000E+00 | 
|sol( 4)| |     4.0000000000E+00 | 
|sol( 5)| |     5.0000000000E+00 | 
|sol( 6)| |     6.0000000000E+00 | 
|sol( 7)| |     7.0000000000E+00 | 
|sol( 8)| |     8.0000000000E+00 | 
|sol( 9)| |     9.0000000000E+00 |

6.3.2 Launching Alain Delon single talkative

 == ESTIMATION OF COMMITED ERROR ==
 Backward error: |A.x-b|/|b| =    8.4657458296985200E-016
 
 == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
          |         Computed (x) | 
_____________________________________________________
|sol( 1)| |     1.0000000000E+00 | 
|sol( 2)| |     2.0000000000E+00 | 
|sol( 3)| |     3.0000000000E+00 | 
|sol( 4)| |     4.0000000000E+00 | 
|sol( 5)| |     5.0000000000E+00 | 
|sol( 6)| |     6.0000000000E+00 | 
|sol( 7)| |     7.0000000000E+00 | 
|sol( 8)| |     8.0000000000E+00 | 
|sol( 9)| |     9.0000000000E+00 |

6.3.3 Launching Alain Delon single thinkative

 == ESTIMATION OF COMMITED ERROR ==
 Backward error: |A.x-b|/|b| =    1.4628721778898641E-016
 
 == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
          |         Computed (x) | 
_____________________________________________________
|sol( 1)| |     1.0000000000E+00 | 
|sol( 2)| |     2.0000000000E+00 | 
|sol( 3)| |     3.0000000000E+00 | 
|sol( 4)| |     4.0000000000E+00 | 
|sol( 5)| |     5.0000000000E+00 | 
|sol( 6)| |     6.0000000000E+00 | 
|sol( 7)| |     7.0000000000E+00 | 
|sol( 8)| |     8.0000000000E+00 | 
|sol( 9)| |     9.0000000000E+00 |

6.3.4 Launching Alain Delon single inverse

 == ESTIMATION OF COMMITED ERROR ==
 Backward error: |A.x-b|/|b| =    5.5572533886510566E-016
 
 == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
          |         Computed (x) | 
_____________________________________________________
|sol( 1)| |     1.0000000000E+00 | 
|sol( 2)| |     2.0000000000E+00 | 
|sol( 3)| |     3.0000000000E+00 | 
|sol( 4)| |     4.0000000000E+00 | 
|sol( 5)| |     5.0000000000E+00 | 
|sol( 6)| |     6.0000000000E+00 | 
|sol( 7)| |     7.0000000000E+00 | 
|sol( 8)| |     8.0000000000E+00 | 
|sol( 9)| |     9.0000000000E+00 |

6.3.5 Launching Alain Delon multiple silent

  1. 2 MPI processes
    [0]  == ESTIMATION OF COMMITED ERROR ==
    [0]  Backward error: |A.x-b|/|b| =    2.0990169310970900E-016
    [0]  
    [0]  == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
    [0]           |         Computed (x) | 
    [0] _____________________________________________________
    [0] |sol( 1)| |     1.0000000000E+00 | 
    [0] |sol( 2)| |     2.0000000000E+00 | 
    [0] |sol( 3)| |     3.0000000000E+00 | 
    [0] |sol( 4)| |     4.0000000000E+00 | 
    [0] |sol( 5)| |     5.0000000000E+00 | 
    [0] |sol( 6)| |     6.0000000000E+00 | 
    [0] |sol( 7)| |     7.0000000000E+00 | 
    [0] |sol( 8)| |     8.0000000000E+00 | 
    [0] |sol( 9)| |     9.0000000000E+00 |
    
  2. 3 MPI processes
    [0]  == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
    [0]           |         Computed (x) | 
    [0] _____________________________________________________
    [0] |sol( 1)| |     1.0000000000E+00 | 
    [0] |sol( 2)| |     2.0000000000E+00 | 
    [0] |sol( 3)| |     3.0000000000E+00 | 
    [0] |sol( 4)| |     4.0000000000E+00 | 
    [0] |sol( 5)| |     5.0000000000E+00 | 
    [0] |sol( 6)| |     6.0000000000E+00 | 
    [0] |sol( 7)| |     7.0000000000E+00 | 
    [0] |sol( 8)| |     8.0000000000E+00 | 
    [0] |sol( 9)| |     9.0000000000E+00 |
    

6.3.6 Launching Alain Delon multiple talkative

[0]  == ESTIMATION OF COMMITED ERROR ==
[0]  Backward error: |A.x-b|/|b| =    2.1195373401239786E-016
[0]  
[0]  == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
[0]           |         Computed (x) | 
[0] _____________________________________________________
[0] |sol( 1)| |     1.0000000000E+00 | 
[0] |sol( 2)| |     2.0000000000E+00 | 
[0] |sol( 3)| |     3.0000000000E+00 | 
[0] |sol( 4)| |     4.0000000000E+00 | 
[0] |sol( 5)| |     5.0000000000E+00 | 
[0] |sol( 6)| |     6.0000000000E+00 | 
[0] |sol( 7)| |     7.0000000000E+00 | 
[0] |sol( 8)| |     8.0000000000E+00 | 
[0] |sol( 9)| |     9.0000000000E+00 |

6.3.7 Launching Alain Delon multiple mixed

[0]  == ESTIMATION OF COMMITED ERROR ==
[0]  Backward error: |A.x-b|/|b| =    2.1113039644227508E-016
[0]  
[0]  == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
[0]           |         Computed (x) | 
[0] _____________________________________________________
[0] |sol( 1)| |     1.0000000000E+00 | 
[0] |sol( 2)| |     2.0000000000E+00 | 
[0] |sol( 3)| |     3.0000000000E+00 | 
[0] |sol( 4)| |     4.0000000000E+00 | 
[0] |sol( 5)| |     5.0000000000E+00 | 
[0] |sol( 6)| |     6.0000000000E+00 | 
[0] |sol( 7)| |     7.0000000000E+00 | 
[0] |sol( 8)| |     8.0000000000E+00 | 
[0] |sol( 9)| |     9.0000000000E+00 |

6.3.8 Launching Alain Delon multiple inverse

[0]  == ESTIMATION OF COMMITED ERROR ==
[0]  Backward error: |A.x-b|/|b| =    1.0884744241440103E-015
[0]  
[0]  == FIRST VALUES OF ESTIMATED VS ANALYTICAL SOLUTION ==
[0]           |         Computed (x) | 
[0] _____________________________________________________
[0] |sol( 1)| |     1.0000000000E+00 | 
[0] |sol( 2)| |     2.0000000000E+00 | 
[0] |sol( 3)| |     3.0000000000E+00 | 
[0] |sol( 4)| |     4.0000000000E+00 | 
[0] |sol( 5)| |     5.0000000000E+00 | 
[0] |sol( 6)| |     6.0000000000E+00 | 
[0] |sol( 7)| |     7.0000000000E+00 | 
[0] |sol( 8)| |     8.0000000000E+00 | 
[0] |sol( 9)| |     9.0000000000E+00 |

Author: HiePACS

Created: 2016-10-12 Wed 13:03

Emacs 24.5.1 (Org mode 8.2.10)

Validate