High-Performance Real-time Architectures for Low-Power Embedded Systems

Risultati finali

The final board: one for the automotive (integrating application-specific hardware), one for the avionic application. Proper exploitation of potential FPGA modules and accelerators available on the boards.

Multi-OS integration and Virtualization

This deliverable resulting from Task 4.3 is a package containing the additional software components for the concurrent execution of Linux and ERIKA Enterprise on the same SoC, and for the synchronization and communication between the two OSs. The deliverable also contains a technical user manual explaining how to install and use the provided package.

Optimized runtime for parallel heterogeneous platforms

The final prototype of the heterogeneous runtime system, plus a report describing the architecture, design principle, main features and results obtained on a first set of benchmarks, in comparison with state-of-the art solutions.

Optimized support for predictability in the accelerator programing environment

The final prototype and a report describing the enhancements implemented for supporting predictability in the target programming model. Support for specification of real-time constraints and interface to RT-analysis tools and environment will be detailed.

Real-time Linux - final

This deliverable is a software package containing the final version of the real-time Linux OS for the reference many-core platform, developed in Tasks 4.1. The deliverable also contains a technical user manual explaining how to install and use the provided package.

Lightweight RTOS - final

This deliverable is a software package containing the final version of the lightweight RTOS for the reference many-core platform, developed in Tasks 4.2. The deliverable also contains a technical user manual explaining how to install and use the provided package.

Detailed Characterization of Platforms

A detailed characterization / model of the performance and energy consumption of the chosen platforms that can be used in the development and evaluation of the schedulers in WP5.

Programming Model(s) analysis and selection

Specification of the chosen programming model(s), including a survey on existing support on the platforms.

Project management plan

Provides detailed guidelines and project management plan. It includes the initial Data Management Plan.

Scheduling Algorithm for Parallel Accelerators

Specification of a scheduling algorithm for execution on accelerators that optimizes accelerator performance and minimizes energy consumption, subject to real-time constraints.

Integrated Schedulability Analysis

Schedulability analysis that considers the combination of the proposed scheduling algorithm executing on the host and on the accelerators.

Year 1 Periodic report

Periodic report after year1. It includes Data Management Report.

Open source results

This deliverable will show the final plan for open source exploitation of the software produced in the project, together with a description of the dissemination activities performed to promote it.

Final exploitation plan

This deliverable will contain the final market analysis for the domains targeted by the project, and will show the results achieved, that is, how the outcomes of HERCULES were/will be pushed in the target market areas. It will also contain the specific exploitation plans by all partners.

Final dissemination report

A final summary of the dissemination activities performed in the project.

Open source strategy plan

This deliverable will be a first draft of the exploitation strategy for the open source software: e.g., how it will be made available, under which license, and possible dissemination activities (presentations, tutorials) for it.

Final Periodic report

Final report of the project. It will include an overview of all outcomes of HERCULES, and it will refer to single final deliverables of all work-packages for details. It includes Data Management Report.

Power-Aware Scheduling Algorithm for Host

Specification of a real-time scheduling algorithm for execution on the host that considers CPU scheduling as well as accesses to shared resources, such as memories, with the objective of minimizing power under real-time constraints.

Specification of Hardware Platforms

Report detailing the results from the platform evaluation and selection for each application (domain).

Application Requirements

This deliverable will list the application requirements and meaningful performance metrics of each application, taking into account the specific application domain (avionics and automotive).

Validation Report

A report of all validation activities on the reference applications running on the specific target architectures, describing the set of experiments performed to validate the use cases under each metric (e.g., energy consumption, performance/cost, predictability), and the improvements achieved with respect to the state-of-the-art and to the initial analysis performed in T1.1, and described in D1.1.

Collaboration and Communication Tools

A report describing all the tools set up to promote and support cooperation among partners.

Pubblicazioni

Energy-aware real-time scheduling in the linux kernel

Autori: Claudio Scordino, Luca Abeni, Juri Lelli
Pubblicato in: Proceedings of the 33rd Annual ACM Symposium on Applied Computing - SAC '18, 2018, Pagina/e 601-608, ISBN 9781-450351911
Editore: ACM Press
DOI: 10.1145/3167132.3167198

SiGAMMA - server based integrated GPU arbitration mechanism for memory accesses

Autori: Nicola Capodieci, Roberto Cavicchioli, Paolo Valente, Marko Bertogna
Pubblicato in: Proceedings of the 25th International Conference on Real-Time Networks and Systems - RTNS '17, 2017, Pagina/e 48-57, ISBN 9781-450352864
Editore: ACM Press
DOI: 10.1145/3139258.3139270

GPUguard: towards supporting a predictable execution model for heterogeneous SoC

Autori: Bjorn Forsberg, Andrea Marongiu, Luca Benini
Pubblicato in: Proceedings of the Conference on Design, Automation & Test in Europe, 2017, Pagina/e 318-321
Editore: European Design and Automation Association
DOI: 10.3929/ethz-b-000222912

HERO: Heterogeneous Embedded Research Platform for Exploring RISC-V Manycore Accelerators on FPGA

Autori: Kurth, Andreas; Vogel, Pirmin; Capotondi, Alessandro; Marongiu, Andrea; Benini, Luca
Pubblicato in: Proceedings of Computer Architecture Research with RISC-V Workshop (CARRV' 17), Numero 11, 2017
Editore: ETH Library
DOI: 10.3929/ethz-b-000219249

Scalable and Efficient Virtual Memory Sharing in Heterogeneous SoCs with TLB Prefetching and MMU-Aware DMA Engine

Autori: Andreas Kurth, Pirmin Vogel, Andrea Marongiu, Luca Benini
Pubblicato in: 2018 IEEE 36th International Conference on Computer Design (ICCD), 2018, Pagina/e 292-300, ISBN 978-1-5386-8477-1
Editore: IEEE
DOI: 10.1109/iccd.2018.00052

Combining PREM compilation and ILP scheduling for high-performance and predictable MPSoC execution

Autori: Joel Matějka, Björn Forsberg, Michal Sojka, Zdeněk Hanzálek, Luca Benini, Andrea Marongiu
Pubblicato in: Proceedings of the 9th International Workshop on Programming Models and Applications for Multicores and Manycores - PMAM'18, 2018, Pagina/e 11-20, ISBN 9781450356459
Editore: ACM Press
DOI: 10.1145/3178442.3178444

Convolutional Neural Networks on Embedded Automotive Platforms: A Qualitative Comparison

Autori: Gianluca Brilli, Paolo Burgio, Marko Bertogna
Pubblicato in: 2018 International Conference on High Performance Computing & Simulation (HPCS), 2018, Pagina/e 496-499, ISBN 978-1-5386-7879-4
Editore: IEEE
DOI: 10.1109/hpcs.2018.00084

On the Cost of Freedom From Interference in Heterogeneous SoCs

Autori: Björn Forsberg, Luca Benini, Andrea Marongiu
Pubblicato in: Proceedings of the 21st International Workshop on Software and Compilers for Embedded Systems - SCOPES '18, 2018, Pagina/e 31-34, ISBN 9781450357807
Editore: ACM Press
DOI: 10.1145/3207719.3207735

HePREM: Enabling predictable GPU execution on heterogeneous SoC

Autori: Bjorn Forsberg, Luca Benini, Andrea Marongiu
Pubblicato in: 2018 Design, Automation & Test in Europe Conference & Exhibition (DATE), 2018, Pagina/e 539-544, ISBN 978-3-9819263-0-9
Editore: IEEE
DOI: 10.23919/date.2018.8342066

The key role of memory in next-generation embedded systems for military application

Autori: Ignacio Sañudo, Paolo Cortimiglia, Luca Miccio, Marco Solieri, Paolo Burgio, Christian Di Biagio, Franco Felici, Giovanni Nuzzo and Marko Bertogna
Pubblicato in: Proceedings of International Conference in Software Engineering for Defence Applications,in Advances in Intelligent Systems and Computing, 2019
Editore: Springer

A Perspective on Safety and Real-Time Issues for GPU Accelerated ADAS

Autori: Ignacio Sanudo Olmedo, Nicola Capodieci, Roberto Cavicchioli
Pubblicato in: IECON 2018 - 44th Annual Conference of the IEEE Industrial Electronics Society, 2018, Pagina/e 4071-4077, ISBN 978-1-5090-6684-1
Editore: IEEE
DOI: 10.1109/iecon.2018.8591540

Deadline-Based Scheduling for GPU with Preemption Support

Autori: Nicola Capodieci, Roberto Cavicchioli, Marko Bertogna, Aingara Paramakuru
Pubblicato in: 2018 IEEE Real-Time Systems Symposium (RTSS), 2018, Pagina/e 119-130, ISBN 978-1-5386-7908-1
Editore: IEEE
DOI: 10.1109/rtss.2018.00021

Work-in-Progress: NVIDIA GPU Scheduling Details in Virtualized Environments

Autori: Nicola Capodieci, Roberto Cavicchioli, Marko Bertogna
Pubblicato in: 2018 International Conference on Embedded Software (EMSOFT), 2018, Pagina/e 1-3, ISBN 978-1-5386-5560-3
Editore: IEEE
DOI: 10.1109/emsoft.2018.8537220

Target following on nano-scale Unmanned Aerial Vehicles

Autori: Daniele Palossi, Jaskirat Singh, Michele Magno, Luca Benini
Pubblicato in: 2017 7th IEEE International Workshop on Advances in Sensors and Interfaces (IWASI), 2017, Pagina/e 170-175, ISBN 978-1-5090-6707-7
Editore: IEEE
DOI: 10.1109/IWASI.2017.7974242

A Software Stack for Next-Generation Automotive Systems on Many-Core Heterogeneous Platforms

Autori: Paolo Burgio, Marko Bertogna, Ignacio Sanudo Olmedo, Paolo Gai, Andrea Marongiu, Michal Sojka
Pubblicato in: 2016 Euromicro Conference on Digital System Design (DSD), 2016, Pagina/e 55-59, ISBN 978-1-5090-2817-7
Editore: IEEE
DOI: 10.1109/DSD.2016.84

On the Accuracy of Near-Optimal CPU-Based Path Planning for UAVs

Autori: Daniele Palossi, Andrea Marongiu, Luca Benini
Pubblicato in: Proceedings of the 20th International Workshop on Software and Compilers for Embedded Systems - SCOPES '17, 2017, Pagina/e 85-88, ISBN 9781-450350396
Editore: ACM Press
DOI: 10.1145/3078659.3079072

Towards predictable execution model on ARM-based heterogeneous platforms

Autori: Premysl Houdek, Michal Sojka, Zdenek Hanzalek
Pubblicato in: 2017 IEEE 26th International Symposium on Industrial Electronics (ISIE), 2017, Pagina/e 1297-1302, ISBN 978-1-5090-1412-5
Editore: IEEE
DOI: 10.1109/ISIE.2017.8001432

Exploring Single Source Shortest Path Parallelization on Shared Memory Accelerators

Autori: Daniele Palossi, Andrea Marongiu
Pubblicato in: Proceedings of the 19th International Workshop on Software and Compilers for Embedded Systems - SCOPES '16, 2016, Pagina/e 197-200, ISBN 9781-450343206
Editore: ACM Press
DOI: 10.1145/2906363.2915925

Ultra low-power visual odometry for nano-scale unmanned aerial vehicles

Autori: Daniele Palossi, Andrea Marongiu, Luca Benini
Pubblicato in: Design, Automation & Test in Europe Conference & Exhibition (DATE), 2017, 2017, Pagina/e 1647-1650, ISBN 978-3-9815370-8-6
Editore: IEEE
DOI: 10.23919/DATE.2017.7927257

Adaptive Coordination in Autonomous Driving: Motivations and Perspectives

Autori: Marko Bertogna, Paolo Burgio, Giacomo Cabri, Nicola Capodieci
Pubblicato in: 2017 IEEE 26th International Conference on Enabling Technologies: Infrastructure for Collaborative Enterprises (WETICE), 2017, Pagina/e 15-17, ISBN 978-1-5386-1759-5
Editore: IEEE
DOI: 10.1109/WETICE.2017.45

Self-Sustainability in Nano Unmanned Aerial Vehicles - A Blimp Case Study

Autori: Daniele Palossi, Andres Gomez, Stefan Draskovic, Kevin Keller, Luca Benini, Lothar Thiele
Pubblicato in: Proceedings of the Computing Frontiers Conference on ZZZ - CF'17, 2017, Pagina/e 79-88, ISBN 9781-450344876
Editore: ACM Press
DOI: 10.1145/3075564.3075580

On the effectiveness of OpenMP teams for cluster-based many-core accelerators

Autori: Alessandro Capotondi, Andrea Marongiu
Pubblicato in: 2016 International Conference on High Performance Computing & Simulation (HPCS), 2016, Pagina/e 667-674, ISBN 978-1-5090-2088-1
Editore: IEEE
DOI: 10.1109/HPCSim.2016.7568399

An energy-efficient parallel algorithm for real-time near-optimal UAV path planning

Autori: Daniele Palossi, Michele Furci, Roberto Naldi, Andrea Marongiu, Lorenzo Marconi, Luca Benini
Pubblicato in: Proceedings of the ACM International Conference on Computing Frontiers - CF '16, 2016, Pagina/e 392-397, ISBN 9781-450341288
Editore: ACM Press
DOI: 10.1145/2903150.2911712

A static scheduling approach to enable safety-critical OpenMP applications

Autori: Alessandra Melani, Maria A. Serrano, Marko Bertogna, Isabella Cerutti, Eduardo Quinones, Giorgio Buttazzo
Pubblicato in: 2017 22nd Asia and South Pacific Design Automation Conference (ASP-DAC), 2017, Pagina/e 659-665, ISBN 978-1-5090-1558-0
Editore: IEEE
DOI: 10.1109/ASPDAC.2017.7858399

Efficient Virtual Memory Sharing via On-Accelerator Page Table Walking in Heterogeneous Embedded SoCs

Autori: Pirmin Vogel, Andreas Kurth, Johannes Weinbuch, Andrea Marongiu, Luca Benini
Pubblicato in: ACM Transactions on Embedded Computing Systems, Numero 16/5s, 2017, Pagina/e 1-19, ISSN 1539-9087
Editore: Association for Computing Machinary, Inc.
DOI: 10.1145/3126560

Multi-Variant Scheduling of Critical Time-Triggered Communication in Incremental Development Process: Application to FlexRay

Autori: Jan Dvorak, Zdenek Hanzalek
Pubblicato in: IEEE Transactions on Vehicular Technology, Numero 68/1, 2019, Pagina/e 155-169, ISSN 0018-9545
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/tvt.2018.2879920

Exploring Shared Virtual Memory for FPGA Accelerators with a Configurable IOMMU

Autori: Pirmin Vogel, Andrea Marongiu, Luca Benini
Pubblicato in: IEEE Transactions on Computers, 2018, Pagina/e 1-1, ISSN 0018-9340
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/tc.2018.2879080

Extending the Lifetime of Nano-Blimps via Dynamic Motor Control

Autori: Daniele Palossi, Andres Gomez, Stefan Draskovic, Andrea Marongiu, Lothar Thiele, Luca Benini
Pubblicato in: Journal of Signal Processing Systems, Numero 91/3-4, 2019, Pagina/e 339-361, ISSN 1939-8018
Editore: Springer Verlag
DOI: 10.1007/s11265-018-1343-1

Combining PREM Compilation and Static Scheduling for High-Performance and Predictable MPSoC Execution

Autori: Joel Matějka, Björn Forsberg, Michal Sojka, Přemysl Šůcha, Luca Benini, Andrea Marongiu, Zdeněk Hanzálek
Pubblicato in: Parallel Computing, 2018, ISSN 0167-8191
Editore: Elsevier BV
DOI: 10.1016/j.parco.2018.11.002

Real-time and energy efficiency in Linux

Autori: Claudio Scordino, Luca Abeni, Juri Lelli
Pubblicato in: ACM SIGAPP Applied Computing Review, Numero 18/4, 2019, Pagina/e 18-30, ISSN 1559-6915
Editore: ACM
DOI: 10.1145/3307624.3307627

A design flow for supporting component-based software development in multiprocessor real-time systems

Autori: Alessandro Biondi, Giorgio Buttazzo, Marko Bertogna
Pubblicato in: Real-Time Systems, Numero 54/4, 2018, Pagina/e 800-829, ISSN 0922-6443
Editore: Kluwer Academic Publishers
DOI: 10.1007/s11241-018-9301-3

Time-Triggered Co-Scheduling of Computation and Communication with Jitter Requirements

Autori: Anna Minaeva, Benny Akesson, Zdenek Hanzalek, Dakshina Dasari
Pubblicato in: IEEE Transactions on Computers, 2017, Pagina/e 1-1, ISSN 0018-9340
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/TC.2017.2722443

GPU-Accelerated Real-Time Path Planning and the Predictable Execution Model

Autori: Björn Forsberg, Daniele Palossi, Andrea Marongiu, Luca Benini
Pubblicato in: Procedia Computer Science, Numero 108, 2017, Pagina/e 2428-2432, ISSN 1877-0509
Editore: ELSEVIER B.V.
DOI: 10.1016/j.procs.2017.05.219

Accelerated Visual Context Classification on a Low-Power Smartwatch

Autori: Francesco Conti, Daniele Palossi, Renzo Andri, Michele Magno, Luca Benini
Pubblicato in: IEEE Transactions on Human-Machine Systems, 2017, Pagina/e 1-12, ISSN 2168-2291
Editore: IEEE Systems, Man, and Cybernetics Society
DOI: 10.1109/THMS.2016.2623482

A software stack for next-generation automotive systems on many-core heterogeneous platforms

Autori: Paolo Burgio, Marko Bertogna, Nicola Capodieci, Roberto Cavicchioli, Michal Sojka, Přemysl Houdek, Andrea Marongiu, Paolo Gai, Claudio Scordino, Bruno Morelli
Pubblicato in: Microprocessors and Microsystems, Numero 52, 2017, Pagina/e 299-311, ISSN 0141-9331
Editore: Elsevier BV
DOI: 10.1016/j.micpro.2017.06.016

Risultati finali

Pubblicazioni

Condividi questa pagina Condividi questa pagina sui social network

Scarica Scarica il contenuto della pagina