HEAD ======= >>>>>>> 36cce3c7378a5c3df4fdb33637490d3507fdb395
We have tested the prototype on a Linux machine running Fedora 29 and Fedora 30. The evaluation code requires access to the /proc
filesystem to read out CPU times, and therefore Linux is required. The code was not tested on Microsoft’s WSL.
The Prototype requires the following libraries to be installed on your system. The build process does not automatically install these. Please use your package manager or follow the links below.
A compiled version of Ticker is provided in the “ticker” directory. Ticker is not required for the prototype to function, but is used in our evaluation as a known state-of-the-art reasoner to test against.
To build the prototype, you will require GHC 8.6 (or higher). It is likely that it will compile against lower versions, but this is untested. The easiest way to build our prototype is by using stack, which can be obtained from https://haskellstack.org.
To build the prototype, run stack build
. If this is your first build using GHC, this may take a long time, because many dependencies will be downloaded and compiled.
To install the resulting executable, run stack install
, which will copy the executable into ~/.local/bin/
. Please make sure that this directory is in your path. If you want to use the executable without installing it, you can use stack run --
. The dashes are required to pass arguments to the prototype rather than stack itself.
The command line interface is largely documented. Use --help
to get help. Specific uses, e.g. for evaluation, are explained in their own sections of this documentation. To run a simulation, execute distributed-sr simulate /path/to/encoding.lars < /path/to/input.log
Examples for encodings and input logs can be found in the test
directory.