HiBench Docker Project

This is the docker image for intel-hibench implementation. We build a pseudo single-node environment on Ubuntu 14.04. Before building/running, please check the file

   $./hibench-docker.conf

for configuration and please set the versions according to your own needs. Otherwise, it will progress by default.

Build and Run Container with HiBench and all dependencies installed

Launch the container by executing:

    $./scripts/run_container.sh cdh

OR

    $./scripts/run_container.sh open-source

The script will launch all related services and run a workload (by default: wordcount) after initialization. When entering the container of the hibench-docker, you are normally guided to a root user, please modify configurations under the HiBench directory :

    #cd ${HIBENCH_HOME}

After running any workload, please check this file

   ${HIBENCH_HOME}/report/hibench.report

for detailed report.

If there is anything wrong with a certain workload, please check this final configuration file:

   ${HIBENCH_HOME}/report/${workload}/${method}/conf/${workload}.conf

Remember to replace the workload variable with the actual one (e.g. wordcount) and the method variable with one of ( mapreduce, spark/scala, spark/java, spark/python ).