Example of how someone might compile the AMRINS code.

Beginning with the 4 tarballs:
-rw-r--r--    1 noel     anag      3522801 May  1 10:31 AMRINS-modifiedBaseline.tar.gz
-rw-r--r--    1 noel     anag      3182261 May  1 10:31 Chombo-modifiedBaseline.tar.gz
-rw-r--r--    1 noel     anag      3534258 May  1 10:32 AMRINS-30April2003.tar.gz
-rw-r--r--    1 noel     anag      3894718 May  1 10:32 Chombo.1.2-30April2003.tar.gz

Assuming the tarballs reside in the current directory, these commands
should do the following: Create a milestoneF directory, populate with
code, then compile serial and parallel for both the modified baseline
and current code.

mkdir -p milestoneF
cd mileStoneF
mkdir -p current
mkdir -p modifiedBaseline
cp ../AMRINS-30April2003.tar.gz current
cp ../Chombo.1.2-30April2003.tar.gz current
cp ../AMRINS-modifiedBaseline.tar.gz modifiedBaseline
cp ../Chombo-modifiedBaseline.tar.gz modifiedBaseline
gunzip */*
cd current
tar -xf AMRINS-30April2003.tar
tar -xf Chombo.1.2-30April2003.tar
cd Chombo/lib
gmake lib DIM=3 DEBUG=FALSE MPI=TRUE 
gmake lib DIM=3 DEBUG=FALSE 
cd ../..
cd dMartin/Chombo-IAMR/_exec-viscous
gmake all DIM=3 DEBUG=FALSE MPI=TRUE
gmake all DIM=3 DEBUG=FALSE 

cd ../../../..
cd modifiedBaseline
tar -xf AMRINS-modifiedBaseline.tar
tar -xf Chombo-modifiedBaseline.tar
cd Chombo/lib
gmake lib DIM=3 DEBUG=FALSE MPI=TRUE
gmake lib DIM=3 DEBUG=FALSE 
cd ../..
cd dMartin/Chombo-IAMR/_exec-viscous
gmake all DIM=3 DEBUG=FALSE MPI=TRUE
gmake all DIM=3 DEBUG=FALSE 

cd ../../../..

ls -l */dMartin/Chombo-IAMR/_exec-viscous/*.ex


the ls should result in something similar to:

-rwxr-x---   1 ndk   lbn   8758416 May  1 14:34 current/dMartin/Chombo-IAMR/_exec-viscous/ns3d.OSF1.g++.f90.MPI.ex*
-rwxr-x---   1 ndk   lbn   7998048 May  1 14:34 current/dMartin/Chombo-IAMR/_exec-viscous/ns3d.OSF1.g++.f90.ex*
-rwxr-x---   1 ndk   lbn   7602848 May  1 14:32 modifiedBaseline/dMartin/Chombo-IAMR/_exec-viscous/ns3d.OSF1.g++.f90.MPI.ex*
-rwxr-x---   1 ndk   lbn   6881648 May  1 14:39 modifiedBaseline/dMartin/Chombo-IAMR/_exec-viscous/ns3d.OSF1.g++.f90.ex*


Example batch script:

---------------------------------------------------------------
#!/bin/csh
#BSUB -J jobname
#BSUB -n 4
#BSUB -W 00:30  
#BSUB -P hp602
#BSUB -q esto_big
#BSUB -o sout
#BSUB -e serr

echo "This job runs on `hostname`."
echo "                                       ---- starting job output"

#  make sure executable, input, and vortexring.dat are around
#  cp ../ns3d.OSF1.g++.f90.MPI.ex .
ls -l *.ex inputs.* ls vortexring.dat
pwd
prun -n 4 ns3d.OSF1.g++.f90.MPI.ex inputs.benchmark.32x32x64

------------------------------------------------------------------

The input files for the current code can be found in the 
same directory as the executable:

halem3% ls -l dMartin/Chombo-IAMR/_exec-viscous/
total 36
drwxr-x---   3 ndk      lbn         8192 May  1 15:15 ./
drwx------  27 ndk      lbn         8192 May  1 15:11 ../
-rw-------   1 ndk      lbn         1483 May  1 15:12 inputs.benchmark.32x32x48
-rw-------   1 ndk      lbn         1483 May  1 15:12 inputs.benchmark.32x32x48.b32
-rw-------   1 ndk      lbn         1482 May  1 15:12 inputs.benchmark.64x64x96
-rw-------   1 ndk      lbn         1482 May  1 15:12 inputs.benchmark.64x64x96.b32
-rw-------   1 ndk      lbn         1488 May  1 15:12 inputs.benchmark.96x96x144
-rw-------   1 ndk      lbn         1488 May  1 15:12 inputs.benchmark.96x96x144.b32

The input files with a ".b32" suffix are for running with a maxboxsize of 32.
Otherwise the maxboxsize is 48.  This is the only difference between the
input files with/without the b32 suffix.


The input files for the modifed baseline are in the same
directory as the executable:

halem3% ls dMartin/Chombo-IAMR/_exec-viscous/inputs.bench*
dMartin/Chombo-IAMR/_exec-viscous/inputs.benchmark.32x32x48
dMartin/Chombo-IAMR/_exec-viscous/inputs.benchmark.64x64x96
dMartin/Chombo-IAMR/_exec-viscous/inputs.benchmark.96x96x144



After running, memory and timing results can be found in the
output file of processor zero (pout.0).

In particular, the following lines:

halem3% grep "AMR Run completed" pout.0
AMR Run completed.  mem= 127.688MB wall-clock time: 821.900 seconds

halem3% grep "Gather end memory" pout.0
Gather end memory from procs:  avg: 130.036  min: 121.832  max: 141.928 (MB)