Category Archives: Uncategorized

Uniserver sponsored research wins an award in ACM Student Research Competition

Konstantinos Tovletoglou, a PhD student at the Institute of Electronics, Communications and Information Technology (ECIT) at Queen’s University Belfast (QUB) won a Bronze Award at the ACM student research competition at the IEEE International Conference on Parallel Architectures and Compilation Techniques (PACT).

The ACM Student Research Competition (SRC), sponsored by Microsoft, offers a unique forum for undergraduate and graduate students to present their original research before a panel of judges and attendees at top-tier international ACM-sponsored and co-sponsored conferences.

Read the full story at this link ….

UniServer at ISPASS 2019

The UniServer project will present two papers at ISPASS 2019 and one of these has been nominated for a best paper award.

The ISPASS conference will be held in Madison Wisconsin March 24-26 2019.

Researchers Paul Whatmough, Shidhartha Das and David Bull, all from project partner ARM and Yiannakis Sazeides and Zacharias Hadjilambrou from project partner University of Cyprus will present their paper

GeST an Automatic Framework for Generating CPU Stress-Tests

The paper, which has been nominated for a best paper award, proposes a framework for automatic stress-test generation.

The second paper presented by UniServer researchers at ISPASS 2019 is

Assessing the Effects of Low Voltage in Branch Prediction Units

The authors of this paper are:

Athanasios Chatzidimitriou, George Papadimitriou,
Dimitris Gizopoulos (University of Athens)
Shrikanth Ganapathy, John Kalamatianos (AMD Research)

The paper explores how low-voltage SRAMS in the Branch prediction unit (BPs + BTB) affect the system operation in terms of power, performance and energy.

Uniserver paper at HPCA 2019

UniServer team members George Papadimitriou, Athanasios Chatzidimitriou, and Dimitris Gizopoulos (University of Athens – UoA) will present a paper

    Adaptive Voltage/Frequency Scaling and Core Allocation for Balanced Energy and Performance on Multicore CPUs

The paper discusses multicore/multithread characterization results on X-Gene 2 and X-Gene 3 systems used in the UniServer project and the effects of core allocation policies for balancing the energy consumption and the performance of the workloads.

It will be presented at HPCA 2019, which is the 25th IEEE International Symposium on High-Performance Computer Architecture that will be held February 16-20 2019, Washington D.C., USA.

Successful UniServer Exploitation Workshop

Members of the Uniserver consortium attended the Exploitation Workshop organized on Monday the 12th of November by Worldsensing and moderated by a team of four experts including VC and IESE professor. The current innovation landscape and entrepreneurship challenges were studied, and a case study developed, before a real-life example was presented, regarding technology transfer. The Value Proposition canvas was produced and will be used as a starting point to develop the exploitation strategy over the remainder of the project. Finally, CEO from various companies shared their experience and answered the questions from the participants in two round tables addressing people and finance topics. The workshop was hosted by the Polytechnic University of Catalonia (UPC), who also introduced the opportunities offered by the Tetramax project (https://www.tetramax.eu/)

UniServer Workshop on Business Strategies for Edge Computing

Worldsensing has organized a business oriented workshop to support two H2020 R&D European projects, mF2C and Uniserver, which work on similar topics – such as edge computing – with the right set of tools to define more appropriate business strategy in-line with the today digital transformation.

The workshop will take place in Barcelona on Monday 12th of November 2018, hosted by the Universitat Politècnica de Catalunya (UPC).

Title: Business tips to understand the consequences of the digital transformation and to adapt today strategy model

List of organizers (click on any name to send an email):

Andrea Bartoli
Denis Guilhot
Ziad Moubarak
Mathieu Carenzo

Worldsensing is a widely recognized global IoT pioneer. It provides customers with the tools to make the right operational decisions based on real-time intelligence. Worldsensing has been awarded the “Premio Inicia for 2017”, as part of the Company of the Year Awards, organized by Spanish newspaper El Periódico.

UniServer at the PACT 2018 Conference

Uniserver was represented at the PACT 2018 conference with two posters for the ACM SRC
student research poster competition:

  • Implementation of a Heterogeneous-Reliability Memory Framework,
    Konstantinos Tovletoglou (Queen’s University Belfast, United Kingdom)
     
  • Harnessing CPU Electromagnetic Emanations for Voltage Noise Characterization,
    Zacharias Hadjilambrou (University of Cyprus, Cyprus)

Series of blog posts from ARM Research on UniServer research

Sidhartha Das from ARM Research has published a three-part blog post in community.arm.com, It highlights work
done in the UniServer project. The posts can be read at the following links

UniServer presented at MICRO 2018

UniServer partner, the University of Athens, presented a 4-hours tutorial on Sunday October 21st in the ACM/IEEE International Symposium on Microarchitecture. The title of the tutorial was “Energy Efficient Computing in Multicore CPUs: Design Margins and Variability” (http://micro51.tutorial.di.uoa.gr/) and the speakers Dimitris Gizopoulos, George Papadimitriou and Athanasios Chatzidimitriou discussed all experiences in characterizing modern multicore ARMv8 CPUs in AppliedMicro’s (now Ampere Computing) machines.

The tutorial was attended by a large crowd of 95 registered people, a strong indication of the importance of energy-efficient computing through margins identification and harnessing and therefore the importance of the findings of UniServer H2020 project.

The following paper was presented at the conference:

Zacharias Hadjilambrou, Shidhartha Das, Marco A. Antoniades, Yiannakis Sazeides, Leveraging CPU Electromagnetic Emanations for Voltage Noise Characterization, 51st Annual IEEE/ACM International Symposium on Microarchitecture (MICRO), October 2018

Here are some pictures for the tutorial session below.

Tutorial by University of Athens at ACM/IEEE MICRO 2018

The University of Athens (UoA) partner in the UniServer project will present a tutorial on Sunday 21 October at the ACM/IEEE MICRO 2018 symposium (ACM/IEEE International Symposium on Microarchitecture) which is taking place in Fukuoka, Japan. More details about the symposium are at this link ….

The title of the tutorial is: Energy Efficient Computing in Multicore CPUs: Design Margins and Variability. Read more at this link ….

The presentation will explain the main challenges of the massive process of characterization and identification of the design margins. The focus will be ARMv8 multicore CPUs (X-Gene 2 and X-Gene 3 of AppliedMicro – now Ampere Computing) voltage margins and variability characterization and corresponding energy savings.

Furthermore, the presenters will discuss accurate and fast margins characterization, system services to expose the hardware behavior to software layers, and modeling in simulators. More details for the contents of the tutorial can be
found here.

Paper from University of Athens at ICCD 2018

The University of Athens (UoA), a partner in the UniServer project, reports that they have had a paper accepted at the IEEE ICCD 2018 symposium (IEEE International Conference on Computer Design) which is taking place in Orlando, FL, USA.

The title of the paper is: Analysis and Characterization of Ultra Low Power Branch Predictors). Authors are A.Chatzidimitriou, G.Papadimitriou, D.Gizopoulos (University of Athens) and S.Ganapathy, J.Kalamatianos (AMD Research).

The paper discusses the performance implications and corresponding energy and power savings when SRAM arrays of modern branch predictors in CPUs are operating at reduced voltage levels.


Copyright © 2020 UniServer Consortium. All rights reserved.