65 lines
2.1 KiB
Plaintext
65 lines
2.1 KiB
Plaintext
To compile the visit plug-in and run in parallel on computational back-end:
|
|
|
|
Checkout code and create a cmake script to compile plug-in.
|
|
Example commands for Rhea:
|
|
|
|
Load the proper modules
|
|
module unload intel gcc
|
|
module load gcc/4.8.2
|
|
module load cmake
|
|
|
|
|
|
Configure cmake to build the plug-in
|
|
cmake \
|
|
-D CMAKE_BUILD_TYPE:STRING=Release \
|
|
-D CMAKE_C_COMPILER:PATH=gcc \
|
|
-D CMAKE_CXX_COMPILER:PATH=g++ \
|
|
-D USE_MPI=false \
|
|
-D USE_CUDA=false \
|
|
-D USE_VISIT=true \
|
|
-D VISIT_ROOT_DIR=/sw/redhat6/visit \
|
|
-D USE_TIMER=false \
|
|
../../LBPM-WIA
|
|
|
|
|
|
Build the visit plug-in
|
|
make install -j 8
|
|
make visit
|
|
|
|
Check that the visit plug-in installed in ~/.visit/2.9.0/linux-x86_64/plugins/databases
|
|
where 2.9.0 is the current version of visit on rhea
|
|
It should have created several shared libraries:
|
|
libELBMDatabase_par.so libELBMDatabase_ser.so libILBMDatabase.so libMLBMDatabase.so
|
|
|
|
Install visit on local machine (2.9.0)
|
|
|
|
Start visit on local machine
|
|
|
|
First time:
|
|
Go to: Options - Host profiles
|
|
Click Remote Profiles - Update
|
|
After list populates, drag host_ornl_rhea.xml to Hosts
|
|
Select ORNL_Rhea from Hosts, click on Machines
|
|
Set account username
|
|
Click on Launch Profiles tab
|
|
Change profile name and timeout if desired
|
|
Click on Parallel tab
|
|
Add Default Bank / Account with project number, set default nodes/processors if desired
|
|
Click Apply, then close window
|
|
Go to: Options - Save Settings
|
|
|
|
Open remote file / launch parallel job
|
|
Click Open file
|
|
Change Host to ORNL_Rhea, enter passcode when prompted
|
|
Open visualization file
|
|
/lustre/atlas/proj-shared/geo106/vis_test/dumps.LBM
|
|
New window will open prompting for # of nodes, processors, Bank (project number) and time limit
|
|
Rhea as 16 physical-cores/node and 64 GB/node
|
|
for example file above I used 1 node, 16 cores
|
|
Visit will call qsub and submit a parallel job that will start when ready
|
|
|
|
Visualize desired quantities
|
|
|
|
|
|
|