Mustafa Ozcelikors and his Mentor Robert Hoettger, have been accepted for participation in this year’s Google Summer of Code. Their project, A4MCAR: A Distributed and Parallel Demonstrator for Eclipse APP4MC will explore different software distribution scenarios along with APP4MC regarding timing and energy efficiency, concurrency issues, and system behaviour on a remote controlled car.
The remote controlled car and its functionality have been part of the Dortmund University of Applied Sciences and Arts research curriculum in conjunction with the AMALTHEA4public project. The A4MCAR graphically presents results of optimization experiments e.g. in terms of APP4MC’s partitioning and mapping configurations. Therefore, the A4MCAR requires a sophisticated hardware and software architecture as well as advanced functionalities through the Raspberry Pi 3 including a touchscreen display, image processing, and sensor aided driving.
Planned software improvements include developing low-level drivers for the XMOS Explorer-Kit multi-core microcontroller and high-level applications for the Raspberry Pi 3 as well as distributing scripts and code snippets to support the community with easier tracing, system management and evaluation options. The project will use the modeling functionality of APP4MC to plan and evaluate affinity-constrained parallel software distributions.
The results will be part of the APP4MC examples downloads, so watch this space for updates! org.eclipse.app4mc.examples.git
On May, the 10 - 11th, APP4MC will be present at the exhibition of the Digital Innovation Forum in Amsterdam. We’ll present our latest platform developments and prototypes as well as some practical demonstrations developed with the APP4MC platform. You can visit us at booth 8 on the exhibition floor Digital Innovation Forum.
SW distribution based on the Polarsys Rover chasis: (https://wiki.eclipse.org/APP4MC/Rover)
EBEAS - Emergency Braking & Evasion Assistance System: Traceability throughout the engineering phases shown with an open source ADAS example
We released a new version of APP4MC with a couple of new features and improvements.
Extract of APP4MC 0.8.0
Model changes (mainly to simplify handling) Improved tooling: model utilities New UI components: Multicore Wizard, Task-Visualizer (draft version)
Model migration support (0.7.2 -> 0.8.0)
Model changes / Improvements
Simplified model handling: Mapping Model, Modes, Time New classifiers to annotate Core and Memory Reworked Property Constraints Extended OS Overheads
On March, the 30th, the talk ‘On Reducing Busy Waiting in AUTOSAR via Task-Release-Delta-based Runnable Reordering’ (2pm) presents analyses regarding resource management based upon APP4MC at the Design, Automation and Test Conference and Exhibition (DATE).
Furthermore, on Friday, March the 31st 10:40am, you can learn how to use APP4MC to analyize and optimize embedded multi-core systems and their performance at parallel2017.
Below, please find the talk’s corresponding paper abstract for the DATE conference.
The increasing amount of innovative software technologies in the automotive domain comes with challenges regarding inevitable distributed multi-core and many-core methodologies. Approaches for general purpose solutions have been studied over decades but do not completely meet the specific constraints (e.g. timing, safety, reliability, affinity, etc.) for AUTOSAR compliant applications. AUTOSAR utilizes a spinlock mechanism in combination with the priority ceiling protocol in order to provide mutually exclusive access to shared resources. The essential disadvantages of spinlocks are unpredictable task response times on the one hand and wasted computation time caused by busy waiting periods on the other hand. In this paper, we propose a concept of task-release-delta-based runnable reordering for the purpose of sequentializing parallel accesses to shared resources, resulting in reduced task response times, improved timing predictability, and increased parallel efficiency respectively. To achieve this, runnables that represent smallest executable program parts in AUTOSAR are reordered based on precedence constraints. Our experiments among industrial use cases show that task response times can be reduced by up to 18,2%.
Based on the APP4MC Eclipse platform, we present the steps required in the development process to make the best use of Multicore controller’s resources. Depending on the development status and the need for analyses, different levels of detail are needed to make targeted evaluations. Examples show which data is required at which development process and what tools can be utilized to extract this data from different sources. The analyzes range from forecast generation in early development phases to the detailed distribution and optimization of the software artifacts to existing hardware.
The 1st official release of APP4MC is published. All IP checks are finally done and the official release review successfully passed. We implemented some new features and enhancements according to project and user given requirements. Further information and entire list of changes can be found in the APP4MC online help https://www.eclipse.org/app4mc/documentation/ or after platform installation
Extract of APP4MC 0.7.2
Model migration support (0.7.1 -> 0.7.2) Improved back references
New core-specific runtimes (instructions) for runnables New model elements for streaming communication (Channels)* New Transmission Policy for communication (e.g. label accesses) New Event-based Stimulus New Data Consistency needs New Counters for InterProcessActivation and EventStimulus Harmonized model elements to describe probability Storage of common elements (e.g. tags) in one central location
Bug fixes / Improvements
Model migration (ISR call graph, Section names, cross document references) Back references (Memory -> Mapping, Scheduler -> CoreAllocation)
New model elements: Physical memory sections, Channels New examples
For details see Release 0.7.2 issues * Visit the APP4MC download page
A new milestone of APP4MC is published. There are new features and enhancements implemented according to project and user given requirements. Further information and entire list of changes can be found in the APP4MC help (after installation).
Extract of APP4MC 0.7.1
To find out more:
View Bugzilla list (https://bugs.eclipse.org/bugs/buglist.cgi?product=App4mc&target_milestone=0.7.1)
Visit the APP4MC download page
Learn more about APP4MC and see it in action at EclipseCon Europe 2016. Explore APP4MC through a demo of a radio-controlled car at the IoT Playground on Wednesday. Start the week by meeting members of the project team at the Unconference.
The APP4MC project team will be participating at this year’s Unconference as part of a meeting of two new Working Groups in the automotive domain, OpenMDM and OpenPASS.
APP4MC in the IoT Playground
Drop by the IoT Playground on Wednesday to explore the potential of the APP4MC platform when used to control a radio-controlled car. Implemented features include visual traffic cone detection and ultrasonic sensor-based obstacle detection. The platform creates models with various configurations and timing behaviors that can be observed and investigated, as well as selected and combined into different features of the car.
The 2nd annual project review held in Bilbao is passed. The project itself presented the work of the last year to show the outcome and usage of the AMALTHEA data model and the APP4MC platform.
Overall the results and feedback of the reviewers were positive. The expectation for the last period of the project is the broader usage and an increasing involvement of outside partners.
To find out more:
Capra, the dedicated traceability management tool is now an official Eclipse EMF project. It allows the creation, management, visualisation, and analysis of trace links within Eclipse. Eclipse Capra is also one of the results of the Amalthea4public project and shares close ties with Eclipse APP4MC.
Capra is a dedicated traceability management tool that allows the creation, management, visualisation, and analysis of trace links within Eclipse. Trace links can be created between arbitrary artefacts, including all EMF model elements, all types of source code files supported by the Eclipse Platform through specialised development tools, tickets and bugs managed by Eclipse Mylyn, and all other artefacts for which an appropriate wrapper is provided. Capra is highly configurable and allows users to create their own traceability meta-model.
To find out more:
Stay tuned for more over the coming weeks!
The Formal Methods for Timing Verification challenge (FMTV) at International Workshop on Analysis Tools and Methodologies for Embedded and Real-time Systems (WATERS) is based on an Amalthea model of a Bosch Engine Control.
About the FMTV Challenge
The purpose of the Formal Methods for Timing Verification (FMTV) challenge is to share ideas, experiences and solutions to a concrete timing verification problem issued from real industrial case studies. It also aims at promoting discussions, closer interactions, cross fertilization of ideas and synergies across the breadth of the real-time research community, as well as attracting industrial practitioners from different domains having a specific interest in timing verification.
The 2016 FMTV Challenge
The 2016 challenge is proposed by Arne Hamann, Simon Kramer, Martin Lukasiewycz and Dirk Ziegenbein from Robert Bosch GmbH.
An initial version of the challenge is available here: FMTV_challenge_2016_Bosch.pdf
The model can be downloaded here: FMTV_Challenge_2016_Bosch_Engine_Control_Model.zip
Summer School at the University of Applied Sciences and Arts in Dortmund from the 27th of June until the 1st of July 2016
More about the Summer School
At EclipseCon France 2016, Toulouse - Capra: A Configurable and Extendable EMF based Traceability Tool
Presentation by Salome Maro, Gothenburg University
At parallel 2016, Heidelberg - APP4MC- eine offene Plattform zur Optimierung der Embedded-Multicore-Performance,
by Harald Mackamul and Joerg Tessmer, Bosch
More about the Parallel Konferenz
Amalthea4Public is partnering with the Eclipse Foundation at Embedded World 2016 in Nuremberg. Drop by the booth to see one of the platform demonstrator projects, an automotive infotainment touch display center.
Article by Christopher Brink and Jan Jatzkowski, University of Paderborn published in the Eclipse newsletter
Last spring, Christopher Brink and Jan Jatzkowski published an overview of the Amalthea4Public project and the roadmap that leads to the APP4MC project.
Back to the top