Verimag

The Language Factories

Overview

The following tools are available for translating other languages into BIP :

More details about the theoretical aspects of these transformations can be found here.


MATLAB/Simulink to BIP

Simulink2BIP is a tool for the translation of Simulink models into the synchronous subset of BIP. The Simulink2BIP can translate part of the discrete-time fragment of Simulink. Simulations are performed using fixed-time step and on single rate mode. Through this page the developer can download the tool as well as documentation about installing, running and debugging the tool. Some examples are also provided.

Distribution

Latest Version : 0.5, Jun. 7, 2010

To use the tool you will need to download it. Below you will find the latest distribution of the tool and the library of the components.

  • The Simulink2BIP tool (download)
  • The Simulink translated library (download)
  • Installation and running instructions (download)

Disclaimer : Simulink2BIP is still in its early development stages. We apologize for any bugs and we cannot guarantee the correctness of the results. The source code is available from the author upon request.

Documentation

Below we provide helpful documents about the tool :

  • Restrictions about the Simulink models to be parsed (download)
  • A report about the translation of Simulink into Synchronous BIP (download)
  • A brief description for future developers of the tool (download)

The developer should be provided also with :

Examples

Here are some examples to download, run and compile :


C to BIP

The C to BIP language factory translates a given C code into a BIP program. It features :

  • the parsing of the entire C code (including an optional pre-processor pass)
  • the generation of one atomic component (i.e. an automaton) for every C functions chosen by the user
  • handles special (user defined list) function calls as inter-component interactions

The factory has been developed so as to be able to use it in various C-related contexts. Currently, it is used :

  • in the MIND project to translate a MIND component architecture into a BIP program
  • in the DOL to BIP factory

The C parsing and manipulations are based on the codegen library, dedicated to this task.


DOL to BIP

dol2bip is a tool for the translation of application software described in the DOL (Distributed Operation Layer, ETHZ) framework into an application software model in BIP. The underlying semantic model of DOL is Kahn process networks. The description consists of two parts :

  • the process network structure, described in XML
  • the process behavior, described in C, one process per file

For each process, there are two particular functions that have to be provided, respectively init and fire. Moreover, processes use specific primitives to communicate (DOL_write/DOL_read to write/read data from the channels). This page provide the links to download the tool as well as documentation about installing and running the tool. Some examples and case studies are also provided.

Distribution

The package is bundled with the standard BIP distribution and hence it also includes the BIP compiler and runtime

System Requirements

  • Linux
  • Java 1.6.

Installing the Toolchain

Extract the archive on your system :

~ $ tar zxf BIP.linux.x86.tar.gz

Point the environment variable BIP2_HOME to this new directory.

for c-shell

~ $ setenv BIP2_HOME $PWD/BIP.linux.x86

for bash

~ $ export BIP2_HOME=$PWD/BIP.linux.x86

The dol2bip translator is invoked by running the $BIP2_HOME/bin/dol2bip executable.

Coding guidelines for process description in C

The C source description for a processes must obey the following :

  • No C++ style variable declaration, e.g., for (int i = 0; i < ...) ... Instead, do int i; for (i = 0; i < ...) ...
  • The structure that defines the local variables (state) of a process must have the predefined name _local_states.

Running an example

Example source structure : Each example folder must contain the application process network description as an XML file, and a src/ folder containing the C files describing the processes (one C file per process).

Example 1 : Generator-Square-Consumer : This example is adapted from the DOL distribution. The figure below illustrates the process network.

The process network in describes in the file example1.xml.

The behavior of processes generator, consumer and square are described respectively in the files generator.c, consumer.c and square.c. The complete source is here.

To translate the DOL specifiation into BIP and execute the BIP model, do the following steps :

  • Step1 : Generate BIP application software model
    ~ $ $BIP2_HOME/bin/dol2bip -f example1.xml
    Creating tmp folder...
    Parsing file : /tmp/DOL2BIP4507904712346052284tmp/libSW.bip
    Generating model : libSW.timod
    Generating model : libSW.model
    Scanning process network ...
    Process name:    generator
    Scanning source: generator.c
    ...............
    Process name:    consumer
    Scanning source: consumer.c
    ...............
    Process name:    square
    Scanning source: square.c
    ...............
    Finished generation of example1_SWModel.bip
    Deleting tmp folder

    The application software BIP model is generated in the file example1_SWModel.bip. This can be compiled and executed using the BIP compiler and runtime as follows :

  • Step2 : Compile
    ~ $ $BIP2_HOME/bin/bipc -f example1_SWModel.bip --genC-execute
    BIP Compiler. Version: frontend: trunk@3019, backend:trunk@3018
    Parsing file : /local/basu/junk/example1/example1_SWModel.bip
    Generating model : example1_SWModel.timod
    Generating model : example1_SWModel.model

    Parser exiting with no errors.

    C code generated
  • Step3 : Build
    ~ $ make -f example1_SWModel.bip.mk
  • Step4 : Execute
    ~ $ ./example1_SWModel.bip.x --quiet
    BIP mono-threaded engine. Version: Engines: trunk@1457
    consumer: 0.000000
    consumer: 1.000000
    consumer: 4.000000
    consumer: 9.000000
    consumer: 16.000000
    consumer: 25.000000
    consumer: 36.000000
    consumer: 49.000000
    consumer: 64.000000
    consumer: 81.000000
    consumer: 100.000000
    consumer: 121.000000
    consumer: 144.000000
    consumer: 169.000000
    consumer: 196.000000
    consumer: 225.000000
    consumer: 256.000000
    consumer: 289.000000
    consumer: 324.000000
    consumer: 361.000000

Case Studies

MJPEG Decoder


Contact | Plan du site | Site réalisé avec SPIP 3.0.22 + AHUNTSIC [CC License]

info visites 671103