ESLAB

Embedded Systems Laboratory

ESLab

The Embedded Systems Lab (ESLab) is specialized in  the design of Dependable Embedded Systems, particularly in the development of hardware and software-based methodologies for the testing, reliability, and security of embedded systems with emphasis on embedded processors and Field Programmable Gate Arrays (FPGAs).

UNIPI_logo

Research topics

Field Programmable Gate Arrays (FPGA) Reliability

  • Radiation effects analysis in COTS FPGAs
  • Single Event Effect (SEE) characterization of COTS FPGAs with radiation and fault injection experiments
  • Reliability analysis of FPGA accelerators (deep neural networks, image compressors, approximate circuits)
  • Single Event Effect (SEE) mitigation approaches
  • Fault Detection, Isolation and Repair (FDIR) methodologies for FPGAs

    • Microprocessor Reliability

  • Functional testing methodologies for embedded processors, multicore processors and SoCs
  • Reliability analysis of microprocessors with radiation and fault injection experiments
  • Fault-tolerant processors for critical applications

    • Hardware Security

  • Security analysis and countermeasure design for FPGAs, embedded systems and ICs
  • Side-Channel & Fault-Injection attacks
  • EDA tools for Hardware Security
  • Physically Unclonable Functions (PUF)

    • news

    Το εργαστήριο χρησιμοποιεί εξοπλισμό του εργαστηρίου Ασφάλειας, που έχει χρηματοδοτηθεί από την Περιφέρεια Αττικής, για την ανάπτυξη κατανεμημένων υπηρεσιών blockchain πολλαπλών κόμβων. Η υποδομή θα χρησιμοποιηθεί για ερευνητικούς και αναπτυξιακούς σκοπούς, σε συνεργασία με ενδιαφερόμενους οργανισμούς του δημόσιου και του ιδιωτικού τομέα.

    AMD incorporates soft error mitigation mechanisms in its UltraScale+ MPSoC devices to recover memory upsets (i.e., faults) before they propagate to the application output and become an error.

    But how effective are these soft error mitigation mechanisms? Can they effectively recover upsets in high altitude or large scale applications under different workloads?

    Our new work entitled “Single Event Effects Assessment of UltraScale+ MPSoC Systems under Atmospheric Radiation”
    (https://lnkd.in/dQjqfDWg) answers the above research questions through a solid study that entails accelerated neutron radiation testing and dependability analysis.

    We test the device on a broad range of workloads, like multi-threaded software used for pose estimation and weather prediction or a software/hardware (SW/HW) codesign image classification application running on the AMD Deep Learning Processing Unit (DPU)

    Η υποδομή έχει χρηματοδοτηθεί από την Περιφέρεια Αττικής στα πλαίσια του επιχειρησιακού προγράμματος «ΑΤΤΙΚΗ» Άξονας Προτεραιότητας: 01 «Ενίσχυση των Μηχανισμών και των Επενδύσεων των ΜΜΕ της Περιφέρειας Αττικής στην Έρευνα και την Καινοτομία» ο οποίος συγχρηματοδοτείται από το Ευρωπαϊκό Ταμείο Περιφερειακής Ανάπτυξης (ΕΤΠΑ) με τίτλο «Ανάπτυξη Υποδομών και Δομών σε κρίσιμες περιοχές και τομείς Ε&Κ σε συμφωνία με την Υλοποίηση της RIS3 της Περιφέρειας Αττικής».

    Approximate Computing Techniques (ACTs) are becoming necessary in many critical Edge computing systems, such as self-driving cars and Earth observation satellites, to increase computational efficiency.
    However, an important question comes to mind when targeting critical systems; Does ACT optimisation negatively affect the reliability of the system, and how can one find optimal design architectures that blend classic mitigation techniques like Triple Modular Redundancy (TMR) with approximation- and precise-based arithmetic hardware units to achieve the best possible computational efficiency without compromising dependability?

    Our work aims to solve this research problem by introducing a Design Space Exploration (DSE) methodology that employs ACTs in arithmetic units of the design and identifies Pareto-optimal microarchitectures that balance all relevant gains of ACTs, such as area, speed, power, failure rate, and precision, by inserting the correct amount of approximation and hardware redundancy in the design.

    Linkedin-in

    Embedded Systems Laboratory

     

    eslab@unipi.gr