Twister2 is a big data tool kit supporting different parallel computing paradigms for big data applications. Twister2 models a parallel computation as a graph. Depending on the requirements the graph can be deployed, executed with different choices making Twister2 capable of running different applications including classic parallel computing, data parallel, streaming or function as a service applications.
The detailed desing of Twister2 can be found at
We use Jira for tracking the progress of twister2.
The google groups mailing list can be used to ask questions and discussions "[email protected]"
The project uses bazel build system. We are using the 0.30 version of bazel.
Compiling the code
bazel build --config=ubuntu twister2/...
Building the packages
bazel build --config=ubuntu //scripts/package:tarpkgs
To run an twister2 application use the twister2 command. Example command
You need to install OpenMPI 3.0 in order to run the following applciation.
./bin/twister2 submit nodesmpi jar examples/libexamples-java.jar edu.iu.dsc.tws.examples.basic.BasicReduceJob