azurehpc/apps/intersect
Cormac Garvey 91e88d43ea Changed TAR to TGZ 2019-10-31 15:58:40 -06:00
..
install_case_intersect_2018.2.sh Changed TAR to TGZ 2019-10-31 15:58:40 -06:00
install_full_intersect_2018.2.sh Changed IX license variable to IX_LICENSE_PORT_IP. 2019-10-31 11:43:11 -06:00
readme.md Corrected readme 2019-10-27 09:50:15 -06:00
run_intersect_2018.2.sh Added APP_INSTALL_DIR and DATA_INSTALL_DIR variables. 2019-10-26 12:13:31 -06:00

readme.md

Intersect installation and running instructions

Prerequisites

Cluster is built with the desired configuration for networking, storage, compute etc. You can see the tutorial or examples folder in this repo for how to set this up.

Dependencies for binary version:

None

Edit intersect full_intersect_2018.2.sh to have the license server and port number, sas url for intersect and eclipse iso Where PORT and IP are port and IP address of license server (e.g 23456@17.20.20.1)

Edit install_case_intersect_2018.2.sh to update sas url for the dataset tar file

Install applications

First copy the apps directory to the cluster. The azhpc-scp can be used to do this:

azhpc-scp -r $azhpc_dir/apps hpcuser@headnode:.

Alternatively you can checkout the azurehpc repository but you will need to update the paths to apps directory according to where you put it.

If you plan on running intersect, you will be required to set-up your own licensing. Intersect will need a valid PORT@IP to your license server.

Intersect installation and running instructions

Install Intersect and eclipse from iso files

azhpc-run -u hpcuser apps/intersect/install_full_intersect_2018.2.sh

Install the data sets for intersect

azhpc-run -a apps/intersect/install_case_intersect_2018.2.sh

Run Intersect

Intersect is run from the headnode (as user hpcuser), First log-in to headnode as user hpcuser.

azhpc-connect -u hpcuser headnode

Next run

qsub -v "casename=<case name>" -l select=2:ncpus=15:mpiprocs=15,place=scatter:excl /apps/intersect/run_intersect_2018.2.sh 

Where "case name" (e.g BO_192_192_28) is the intersect case you want to run)

To see if the job is running do

qstat -aw

Install and run intersect Benchmarks using Azure CycleCloud Cluster

Prerequisites

These steps require a Azure CycleCloud cluster with PBS. The cyclecloud_simple_pbs template in the examples directory a suitable choice.

Follow the steps in the examples/cyclecloud_simple_pbs/readme.md to setup cycle, import the template and start cluster.

Log in to the headnode of the cluster (from cycleserver):

    $ cyclecloud connect master -c <cyclecloud cluster name>

Installing Intersect

You will need to copy the apps/intersect folder to the cyclecloud master.

Run the following to install ntersect on the cluster (in /scratch):

export APP_INSTALL_DIR=/scratch

apps/intersect/install_full_intersect_2018.2.sh

Install the data sets for intersect

export DATA_INSTALL_DIR=/scratch

apps/intersect/install_case_intersect_2018.2.sh

Running Intersect

Copy apps/intersect to the cyclecloud master node.

To run on two HB nodes with 8 total cores (4 cores on each node) run ( Intersect installation and case model are in /scratch)

qsub -l select=2:ncpus=60:mpiprocs=4 -v case=BO_192_192_28,APP_INSTALL_DIR=/scratch,DATA_INSTALL_DIR=/scratch apps/intersect/run_intersect_2018.2.sh