october 2020 - hipeac
TRANSCRIPT
61
Future Real-Time Safety-Critical RISC-V Computers
A holistic approach for Cyber-Physical Systems of Systems
Welcome to the virtual Computing Systems Week
Join
the HiPEAC
virtual
events
OCTOBER 2020
Innovation Europe The impact of COVID-19 on the HiPEAC Jobs Portal
EPEEC Programming Guidelines for Parallel Applications
4 11 12
contents
3 WelcomeKoen De Bosschere
4 Innovation Europe First Milestones towards Future Real-Time Safety-Critical
RISC-V Computers
6 Innovation Europe AMPERE: A holistic approach for real-time, high-
performance and energy-efficient Cyber-Physical Systems of Systems
8 Innovation Europe SMART4ALL is a four-year innovation action promoting
marketplace-as-a-service in South East Europe
10 Innovation Europe HENSOLDT Cyber presents MiG-V, the first RISC-V
Processor “Made in Germany“ for Security Applications
11 Peac Performance EPEEC Programming Guidelines for Parallel Applications -
Increasing application developers’ productivity
12 HiPEAC futures Impact of Covid-19 on HiPEAC Jobs Portal
14 HiPEAC futures Acaces 2020 Career Session
16 HiPEAC futures Three-minute thesis
17 News Simulating COVID-19 and flu spread using HiDALGO
Or how technology can support decision-making for an effective response to a pandemic
17 News CPS Convergence and Society
18 News Welcome to the virtual Computing Systems Week hosted
by Tampere University
19 News Afterthoughs on the ACACES2021 virtual summer school
20 HiPEAC Conference 18-20 January 2021, Budapets
HiPEAC is the European network on high performance embedded architecture and compilation.
hipeac.net @hipeac hipeac.net/linkedin
HiPEAC has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement no. 871174.
Cover image: Laura Vanzo, Visit Tampere
Design: www.magelaan.be
Editor: Rebecca Gorby & Vicky Wandels
Email: [email protected]
HiPEACINFO 612
Afterthoughs on the ACACES2021 virtual summer school
Virtual Computing Systems Week hosted by Tampere University
Acaces 2020 Career Session
14 1918The word of the year 2020 will probably be COVID-19. It started like yet another
virus, but after some time, it became clear that it was the so-called “big one”. COVID-
19 became a humbling wake-up call. We sometimes believe that we are the masters of
the universe, but we are not. A particle of 100 nm was, in no time, able to change the
lifestyle of seven billion people on planet Earth, caused more than one million of deaths,
ended thousands of businesses, and even changed the course of elections. The impact
of the virus is on a par with the impact of a major natural disaster. Eventually, we will
outsmart the virus with a vaccine, but it will take at least another year before the virus
is under control. Inoculating seven billion people (twice) is no small task. Even at a rate
of one shot per minute, it will take more than 100 000 person years, and depending
on immunity response, this process might have to be repeated regularly. This means
that unfortunately, COVID-19-related issues will not disappear the day after the first
vaccines become widely available.
By the time we will have acquired enough immunity, the world will have changed:
new leaders will be in power, new businesses will have emerged, millions of people will
have changed jobs, and the world economy will have evolved. Now is the right time
to think outside the box and to propose bold revisions, to work towards a sustainable
economy, to reduce inequality and injustice, to remedy the excesses of globalization and
to create a better society. This is a once-in-a-lifetime opportunity for visionary leaders.
I hope they will rise soon.
It is crucial that we learn lessons from this pandemic: global problems can only be
solved through solid science and global collaboration. COVID-19 is just one example of
a global problem. Climate change is another, as is the reduction of biodiversity. COVID-
19 teaches us that there are no simple solutions for complex problems and that we need
the joint creativity (science, politics, economy, media, education, etc.) of the whole
world to find solutions for such problems. Fortunately, complex solutions will always
require advanced computing, and that means that the computing industry will have a
bright future, and will become even more relevant for society than it already is. I hope
that we are all looking for new business opportunities in these turbulent times.
Take care,
Koen De Bosschere, HiPEAC coordinator
welcome
HiPEACINFO 61 3
FIRST MILESTONES TOWARDS FUTURE REAL-TIME SAFETY-CRITICAL RISC-V COMPUTERS
Innovation Europe
The space industry, like other industries that make use of
microelectronic circuits, depends to a great extent on American
technology. Historically, the United States has had a lead
in the sector by delivering state-of-the-art technologies. An
example of this is the GAIA (Global Astrometric Interferometer
for Astrophysics) mission that was forced to use American
technology due to a lack of an European alternative. This
scenario can be problematic for European actors when the
United States is able to exclude other players from the use of its
technology, as well as to prevent licensing.
In recent years, European developments within space-
computing have progressed and American space missions also
currently use European technology for on-board processing. As
the space industry progresses to more advanced system-on-chip
architectures, the current lead manufacturers for space-grade
processors in terms of performance is the United States again.
At the same time, the trend towards higher integration does not
only require high-performance. There is an increasing need for
cybersecurity and partitioning as multiple software functions
that were previously run on separate components in isolation
are moved to the same system-on-chip.
Under such conditions, the De-RISC project was born. The
project aims at productizing the first market-ready European
Cooperation for Space Standardization (ECSS) level B space-
certifiable software-hardware platform using the RISC-V
open-standard Instruction Set Architecture (ISA) and entirely
developed by European companies. Furthermore, the project
also aims at providing avionics grade readiness, seeking
DO-178C level B certification and even anticipating its usage in
other critical domains, like automotive or railways.
Currently, the project has been running for about one year,
and it is already achieving its first milestones. Platform
requirements have been fully defined and early versions for
Field Programmable Gate Array (FPGA) of the integrated
multicore with enhanced quota and performance monitoring
units are being tested and further developed. It is expected
that by the end of the project, in 2022, the platform will reach
readiness level TRL8.
The De-RISC platform rests on two fundamental pillars: high-
performance and safety. Performance is delivered by a 64-bit
multicore architecture composed of a single scalable General-
Purpose Processor (GPP), which in turn is comprised of four
NOEL-V cores (i.e. Cobham Gaisler’s RISC-V version of their
insignia processor, the LEON-5). Additionally, three levels of
cache (core-private, GPP-shared and system-shared) and new
processor cores with fully pipelined IEEE-754 Floating point
units allow the platform to provide an increase in computational
performance, compared to current systems.
European technologies for a better-integrated world
HiPEACINFO 614
Innovation Europe
monitoring units, will increase the amount of available tools to
system integrators and platform users by providing fine-grain
evidence of correct timing behavior.
The platform is required to have a critical configuration
that permits the user to either eliminate or bound temporal
interference, and where core-local activities are not affected by
another core’s activity. An example of this is the case of the
Level 2 cache, which will allow partitioning across cores to
avoid data evictions to critical tasks. This critical configuration
also allows the user to disable cache coherency (maintained
within the GPP) for certain memory areas without shared data
in situations where cores mutually invalidate cache lines, hence
making it difficult to predict latencies.
As the development of the platform continues, more features
tackling security concerns and resource partition will be
implemented to reach platform readiness.
The project is composed by Cobham Gaisler (Sweden), fentISS (Spain),
Thales Research & Technology (France) and the Barcelona
Supercomputing Center (BSC) (Spain), each partner contributing
crucially to the project’s success. Gaisler is in charge of developing the
hardware platform, while fentISS develops the software stack. BSC adds
to the platform with enhanced hardware safety features and Thales
validates the design with space use cases, like a telemetry and
telecommand application, which provides a wide spectrum of functional
testing.
PROJECT NAME: De-RISC: Dependable Real-time Infrastructure for
Safety-critical Computer
START/END DATE: 01/10/2019 – 31/03/2022
KEY THEMES: RISC-V, System-on-Chip
PARTNERS: fentISS, BSC, Thales Research and Technology,
Cobham Gaisler
BUDGET: € 3,444,625
www.derisc-project.eu
Functionality-wise, the platform will support plenty of well-
known standard Input/Output (IO) peripherals (e.g. I²C, SPI,
2x CAN-FD interfaces, 2x SpaceWire interfaces, 2x SpaceFibre
interfaces) allowing its applicability to other domains beyond
space. The system-on-chip platform also provides a fault-
tolerant DDR2 and DDR3 controller with strong error detection
and correction capabilities.
On the safety side, De-RISC combines state-of-the-art software
and hardware solutions to control the added high-performance
features. The XtratuM hypervisor, together with XRE (XtratuM
Run-time Environment, for bare metal applications) and LithOS
operating system, all of them developed by fentISS, will fully
support the platform, which will be ported from their current
SPARC implementations, and will provide temporal and
spatial isolation to critical applications, hence enforcing safety
measures. Additionally, the inclusion of enhanced observability
and controllability mechanisms, like advanced performance
Image 1. Detail from De-RISC's architecture showcasing a scalable
General-Purpose Processor (GPP) element composed of RISC-V cores.
De-RISC's platform will feature one GPP with four cores.
Image 2. XtratuM with XRE and LithOS running on top of the
De-RISC hardware platform.
HiPEACINFO 61 5
Innovation Europe
Cyber-Physical Systems of Systems (CPSoS) are enabling
technologies to address societal and industrial challenges
across multiple application domains like autonomous and
safe mobility as well as smart manufacturing and sustainable
production, among others. The applications running in these
systems require not only high levels of performance, but also the
fulfillment of non-functional requirements like dependability,
real-time response, resiliency, safety and energy-efficiency.
Although highly parallel heterogeneous embedded architectures
like Digital Signal Processor (DSP) fabrics and Field
Programmable Gate Array (FPGA) accelerators can provide the
performance capabilities needed in CPSoS, these are commonly
developed using model-driven technologies that facilitate the
description of the system, but lack the mechanisms to effectively
describe parallelism and certain non-functional requirements.
Furthermore, parallel programming models also leave non-
functional requirements aside.
The AMPERE project is developing a new generation of
software programming environments for low-energy and highly
parallel and heterogeneous computing architectures, capable of
implementing correct-by-construction advanced Cyber Physical
Systems (CPS).
The key innovation of the AMPERE software architecture will be
its capability of transforming the system model description of
the CPS based on specific model-driven languages to the parallel
programming models supported by the underlying parallel
architecture. Moreover, the AMPERE software architecture will
fulfill the non-functional requirements (i.e., real-time, safety,
energy-efficiency, security, reliability) imposed due to the cyber-
physical interactions and captured in the system description.
Figure 1 shows the components of the AMPERE software
architecture and their flow, described next:
• Domain Specific Model Language (DSML). Enhanced model-
driven languages that describe non-functional constraints
(e.g., performance, energy, time, etc.) in the context of
parallel heterogeneous computing.
• Code synthesis tools, compilers and analysis tools. Tools
capable of extracting control- and data-flow information,
as well as transforming the code to maximize multi-criteria
optimization and correctness.
• Runtime libraries. In charge of: (1) orchestrating parallel
execution according to the model, (2) efficiently managing
offloading to accelerator devices, and (3) supporting efficient
accelerator execution while preserving non-functional
requirements.
• Operating system and hypervisor. Support a variety of
architectures and accelerators, as well as safe and secure real-
time management of hardware resources.
AMPERE will be tested in two use cases:
• Automotive: Intelligent Predictive Cruise Control (PCC). This
use case, provided by Bosch, is an example for the increasingly
autonomous decision-making capabilities of advanced
automotive systems. It is composed of four components:
Adaptive Cruise Control (ACC), the powertrain control
subsystem, the advanced PCC and Traffic Sign Recognition
(TSR) subsystems. The PCC extends the ACC functionalities
using data from the electronic horizon (topographical data
like curvature, inclines or speed limits) to provide information
about the route ahead. The deep learning TSR detects road
signs and provides limits to the PCC, considering, for example,
construction sites.
AMPERE: A HOLISTIC APPROACH FOR REAL-TIME, HIGH-PERFORMANCE AND ENERGY-EFFICIENT CYBER-PHYSICAL SYSTEMS OF SYSTEMS
Figure 1. AMPERE software development ecosystem.
HiPEACINFO 616
Innovation Europe
PROJECT NAME: AMPERE: A Model-driven development framework for
highly Parallel and EneRgy-Efficient computation supporting
multi-criteria optimization.
START/END DATE: 1/01/2020 – 31/12/2022
KEY THEMES: Computing technologies and engineering methods for
Cyber-Physical Systems of Systems
PARTNERS: Barcelona Supercomputing Center (BSC) (coordinator),
Instituto Superior de Engenharia do Porto (ISEP), ETH Zürich (ETHZ),
Scuola Superiore Sant’Anna (SSSA), Evidence SRL, Bosch GMBH,
Thales, Thales Italy SPA and SYSGO SRO.
BUDGET: € 4.9M.
WEBSITE: ampere-euproject.eu
• Railway: Obstacle Detection and Avoidance System (ODAS).
This use case, provided by Thales Italy, consists of a real
demonstrator on the Florence Tramway Network implementing
and ODAS supporting the tram driver and improving the
level of safety of the transportation system. It incorporates
two main subsystems, the Sensor Data Fusion (SDF) and
the AI Analytics (AI) components. The SDF collects a large
mass of raw data from multiple advanced sensors, while the
AI components incorporate machine learning (e.g., SVM) and
deep learning (e.g., CNN) algorithms to identify and track
objects, and extract information to be displayed to the tram
driver.
The AMPERE team during the first meeting of the project in Barcelona, Spain.
HiPEACINFO 61 7
Innovation Europe
The goal of the project is to build capacity among European
stakeholders via the development of self-sustained, cross-border
experiments that transfer knowledge and technology between
academia and industry.
SMART4ALL Management Team: Nikolaos Voros, Project
Coordinator (University of Peloponese, Greece) – Michael
Huebner, Project Sub-coordinator (Brandenburg University
of Technology, Germany) – Christos Antonopoulos, Technical
Manager (University of Peloponese, Greece) – Georgios
Keramidas, Tecnical Manager (Aristotle University of
Thessaloniki, Greece)
The project targets Customized Low Energy Computing (CLEC)
in the Cyber-Physical (CPS) and the Internet of Things (IoT)
domains by combining a set of unique characteristics that join
together i) different cultures, ii) different policies, iii) different
geographical areas and vi) different application domains. The
SMART4ALL vision will be realized mainly through funded
Pathfinder Application Experiments (PAEs) that will enable the
transformation of academic knowledge into industry, especially
targeting South Eastern European countries that are currently
underrepresented in European funding instruments. In this way,
community building, strategy development, and ecosystem
learning are envisioned for boosting high-quality research and
development in South Eastern Europe (SEE).
SMART4ALL IS A FOUR-YEAR INNOVATION ACTION PROMOTING MARKETPLACE-AS-A-SERVICE IN SOUTH EAST EUROPE
Digited Anything
Digited Transportation
Digited Agriculture
Digited EnviromentSMARTALL vision
Thematic Pillars of SMART4ALL
HiPEACINFO 618
Innovation Europe
In this direction, SMART4ALL knowledge and technology
transfer experiments are expected to improve the level of
support in SEE, by providing better services through the
creation of local DIHs or representative nodes of already
existing European DIHs. The goal is to reveal new examples
and best practices with high potential of becoming market
success stories either at the local or the regional level. In
addition, SMART4ALL will help to identify efficient methods
for discovering and encouraging SMEs that will take advantage
of the proposed digital development process.
Innovative Market PlaceSMART4ALL offers a unique concept called Marketplace-as-a-
Service (MaaS). MaaS is the key differentiator of SMART4ALL
from existing approaches since it reduces the development
effort, e.g. to move from an idea to a prototype. SMART4ALL
MaaS includes cloud services, related platforms, tools and
middleware frameworks, and design service facilities mainly
focusing on open-source technologies. Moreover, services
(customized to the four thematic pillars of the project) are
included in the MaaS e.g., personalized links to relevant
events, customized web pages, and matchmaking (technology
suppliers-technology receivers) activities.
Open Calls StructureSMART4ALL has defined a robust process to choose up to 88
PAEs by launching three Types and nine Open Calls addressing
the four main prioritized verticals.
There will be three types of PAEs:
- KTEs – Knowledge Transfer Experiments: A novel type of
experiment that allows smaller projects, or less mature ideas,
to be presented and tested. KTEs act as internships and short-
term training programmes for unemployed people for vacant
digital jobs (expected duration: three months, budget: up to 8,000€).
- FTTEs – Focused Technology Transfer Experiments: Within
these types of experiments, one party transfers to the
receiving partner a specific hardware (HW) or software (SW)
technology in order to enable improved products or processes
(expected duration: up to nine months, budget: up to 80,000€).
- CTTEs – Cross-domain Technology Transfer Experiments:
Complex multidisciplinary transfers and productization of
novel CLEC technologies to wider markets (expected duration: 12 months, budget: up to 80,000 €).
Cut off Dates
SMART4ALL will organize nine open calls and three per each type
of calls: https://smart4all-project.eu/opencalls-apply-now/
SMART4ALL offers funding of 2,2 million Euros via nine open calls and
novel coaching services from leading experts in ethics, technology,
funding and business development. It will support 88 cross-border
pathfinder application experiments (PAEs) from European consortia.
https://smart4all-project.eu/
SMART4ALL Market as a Service Conceptual Architecture
SMART4ALL process for 3rd parties’ PAEs selection and execution
HiPEACINFO 61 9
Innovation Europe
HENSOLDT CYBER PRESENTS MIG-V, THE FIRST RISC-V PROCESSOR “MADE IN GERMANY“ FOR SECURITY APPLICATIONS
HENSOLDT Cyber GmbH, a developer of highly secure
embedded systems for the IT market, announced that it
successfully designed and produced the first RISC-V processor
“Made in Germany”. Named MiG-V, it addresses the security
needs of connectivity applications in areas like the Internet of
Things.
It enables customers to build connectivity-oriented security
applications in areas like the Internet of Things.
MiG-V is protected against malicious manipulations during
design and manufacturing via advanced logic encryption. This
way, Kill Switch threats like hardware Trojans are excluded,
creating a secure solution for gateways, interfaces to the
external world, and other products.
The core of the MiG-V is based on the CV64, an open-source
RISC-V core developed by ETH Zurich, originally named Ariane.
It is implemented as an RV64IMAC with a 64-bit integer CPU
and standard extensions for integer multiplication and division
(M), atomic (A) and compressed instructions (C). It is produced
using a 55nm CMOS process. The design was supported by Chief
Scientist Prof. Rainer Leupers from RWTH Aachen University.
Connectivity Oriented PeripheralsThe MiG-V system-on-chip integrates 1 MB of internal SRAM,
2 MB Flash memory and an SDRAM controller with a clock
speed of up to 100 MHz. Communication interfaces include two
10/100 Mbps Ethernet MAC controllers, one QSPI and three
SPI controllers with up to 30 MHz and one SPI slave interface
with up to 40 MHz, as well as three UART controllers and one
I2C controller. The chip operates at 3.3V supply voltage.
“MiG-V is a milestone in the development of general-purpose
embedded processors, because it helps to create a secure IT
instead of IT security” says Sascha Kegreiß, CTO at HENSOLDT
Cyber. “Together with TRENTOS-M, our seL4 microkernel-
based operating system, customers can design systems in
various areas with a built-in security level hardly ever achieved
before” Kegreiß adds.
Initial tests were completed successfully, and further integration
and detailed evaluations are underway. Once completed, the
start of the mass production of the MiG-V processor will be
scheduled.
For more information about HENSOLDT Cyber’s MiG-V:
www.hensoldt-cyber.com/mig-v
About HENSOLDT Cyber
Founded in 2017, HENSOLDT Cyber GmbH is a German company based
in Taufkirchen near Munich that develops embedded information
technology products meeting the highest security requirements. These
integrate a highly secure operating system with security hardened
hardware, thus creating a secure IT instead of IT security for the global
IT market. The company combines more than 50 years of experience in
defense and security electronics of the HENSOLDT Group with world-
class expertise in hardware and software development. HENSOLDT
Cyber currently employs around 40 people at various locations.
Further information about the company can be found at
www.hensoldt-cyber.com
HENSOLDT Cyber MiG-V processor
Marian Rachow, CEO - Sascha Kegreiß, CTO
HiPEACINFO 6110
Peac Performance
These recommendations contribute to
the project’s goals by establishing a clear
path to get application codes suited
for parallelization and exploitation
of heterogeneous resources. They are
relevant to application developers
interested in high productivity, as their
aim is to facilitate the compilers’ work and
improve the applications’ performance in
terms of execution time.
“The EPEEC guidelines will allow to
increase the programmer’s productivity
while developing modern applications”
said Antonio J Peña, Coordinator of
EPEEC and Lead of the Accelerators
and Communications for HPC Team at
Barcelona Supercomputing Center (BSC).
“The publication of the EPEEC guidelines is
an important milestone for the application
developers because it will guarantee a
smooth transition from the sequential
version or poorly efficient parallel version
of each code to a highly scalable version
ready to leverage the capabilities of
exascale-class heterogeneous parallel
hardware”, said Stéphane Lanteri, EPEEC
Principal Investigator (INRIA).
The document presents a high-productivity
approach for the development of high-
performance applications based on parallel
programming best practices used by
expert developers in the High Performance
Computing (HPC) community. It introduces
the EPEEC methodological framework,
which splits the parallelization process into
three steps:
1. Prepare the code for parallelism: to
code the application in such a way
that reduces the cost/effort of parallel
software development and maintenance.
2. Create a first parallel version of your
code: to develop a parallel version of
existing sequential code that runs faster.
3. Optimize your parallel code: to fine-
tune the parallel code to obtain peak
performance of the target hardware
platform.
In addition to high-productivity and
the development of a high-performance
programming environment, EPEEC’s goals
include the efficient and energy-aware
management of hardware heterogeneity,
both in terms of processing elements and
memory subsystems, further favouring
coding productivity.
The EPEEC Programming Guidelines for Parallel
Applications are available on the project’s
website: https://epeec-project.eu/results/
programming-guidelines
EPEEC Programming Guidelines for Parallel Applications -Increasing application developers’ productivity
The work towards a highly productive programming environment for heterogeneous exascale computing carried out by the European project EPEEC has reached an important milestone with the publication of the EPEEC Programming Guidelines for Parallel Applications.
EPEEC’s methodological framework for parallelization
HiPEACINFO 61 11
The restrictions that were implemented are having a severe impact1 on
economies and the labour market all across the European Union (EU);
however, the outcomes in some countries may vary, because of their
specific characteristics, economic structures and institutions. This will
result in unequal impacts in terms of the overall job market.
Although the previsions are pessimistic for the next few months, current
employment statistics haven’t yet been affected by the gloomy outlook,
with the EU reporting only a 0,5% increase in the unemployment rate
compared to December 2019 (up to 6.7% in June, from 6.2%). This can
be explained by the containment measures and aid packages provided
by the EU to support the governments, as well as some legal issues of
what constitutes an “unemployed person” in Europe, as per the
following2,3. The European Commission expects a shrinking of Gross
Domestic Product (GDP) in the Union of about -8,7% by the end of the
year, but a growth of 6.1% in 2021.4
The HiPEAC Jobs Portal (https://www.hipeac.net/jobs/#/), though it
focuses on a less vulnerable field of work– Information and Systems
and Computer Engineering – when compared to others like Tourism,
Retail and Hospitality, has also been affected by the pandemic. As you
can see in Figure 1, the number of jobs posted this year, when compared
to 2019, has decreased by 50 vacancies - especially remarkable during
March and April- the months when Europe was most affected by COVID-
19, which resulted in the lockdown of many of the European countries.
Also, in terms of traffic to our jobs portal, the number of visitors has
decreased notably when compared to previous year(s) (see Figure 2).
An interesting comparison that can be also be made is between the
different types of employers using the portal – academia, for universities
and government institutions, and industry or small and medium-sized
enterprises (SME´s) for privately owned companies – there is a 30%
reduction in the private sector job openings in 2020 when compared to
the previous year, indicating that the private sector could be more
affected by the pandemic, and subsequently has been more conservative
with hiring policies and less open to publish more positions during this
period. It also indicates that industry is more linked to physical events
and tradeshows compared with academia, as major events such as
DATE and CSW were either moved online or postponed until after the
lockdown period (Figure 3).
Nonetheless, there is hope of a recovery as the institutions, companies
and overall job market adapt themselves to this new reality.
The impact of COVID-19 on the HiPEAC Jobs PortalIn a seemingly connected and globalized world, the
impact of the COVID-19 pandemic on society is undeniable. From travel restrictions to schools,
shopping malls, offices and major companies shutting down, the pandemic has affected everyone
around the globe – postponing the free circulation of people and affecting, of course, the European
economy and job market.
HiPEAC futures
HiPEACINFO 6112
The implementation of the HiPEAC Internship Programme – now
with remote/virtual opportunities (https://www.hipeac.net/jobs/#/
characteristics. And all in all, it is still too early to see the longer-term
effect.
HiPEAC continues readjusting itself, as new features are incorporated,
such as the possibility to incorporate career-related videos5. HiPEAC
will keep innovating and remodeling its activities in order to overcome
this downturn, as we are fully committed to provide the best possible
service to our community. Our support to the partners in the network
will only increase over the course of the year – with tailor-made support
for new job postings, as well as with help with both the promotion and
the development of these opportunities. For any support needed,
contact [email protected]
Figures and charts
The impact of COVID-19 on the HiPEAC Jobs Portal
HiPEAC futures
1 The impact of COVID-19 confinement measures on the EU labour market https://ec.europa.eu/
jrc/sites/jrcsh/files/jrc.120585_policy.brief_impact.of_.covid-19.on_.eu-labour.market.pdf]2 https://ec.europa.eu/eurostat/documents/2995521/10159296/3-30012020-AP-EN.pdf/
b9a98100-6917-c3ea-a544-ce288ac096753 https://ec.europa.eu/eurostat/statistics-explained/index.php/Unemployment_statistics4 https://ec.europa.eu/commission/presscorner/detail/en/ip_20_1269 - Summer 2020
Economic Forecast5 https://www.hipeac.net/news/6927/career-job-related-videos-now-available-at-hipeac/
Reference taken from the OECD report of worker security and the
COVID-19 crisis at http://www.oecd.org/employment-outlook/#report)
DON’T MISS OUR UPCOMING ACTIVITIES
BY FOLLOWING US:
hipeac.net/linkedin
@hipeacjobs /@hipeac
hipeac.net/jobs
Figure 1: Number of jobs posted on the HiPEAC jobs portal
Figure 2: Weekly number of page views of the HiPEAC jobs portal
HiPEACINFO 61 13
ACACES 2020 CAREER SESSIONAn overview of professional perspectives with different career backgrounds
Introduction: On 15 July, we invited Andrea Kells from Arm, Frank K. Gürkaynak from ETH Zürich, and Ray Garcia from Buoyant Capital to share with our
students different perspectives on career paths, in a virtual career session at ACACES. From industry,
academia and at the entrepreneurial level, we had the opportunity to hear first-hand from these
talented professionals their opinions on setting up a career strategy.
HiPEAC futures
From the academic perspective, we learned that focusing on
research allows you to work as a professional and specialize in a
certain field.
Frank told us that “University is an opportunity, as you'll be
surrounded by a lot of interesting people and you have the chance
to learn a lot. For this specific field of work, you have to understand
what you are doing and why are you doing it that way, as that is the
only way to achieve academic results.”
He also said that, as a teacher, many students often come to his
office to discuss grades – but in the end, they are only feedback for
the student – low grades mean that you should “work more” while
good grades mean that “you are getting it”.
ACADEMIC“For this specific field of work,
you have to understand what
you are doing and why are you
doing it that way, as that is
the only way to achieve
academic results.”
14 HiPEACINFO 61
HiPEAC futures
The entrepreneurial perspective by Ray Garcia is, of course, a very
independent approach – each person must define their own
definition of failure and success to build their career strategy. This
will change from individual to individual and both failure and
success are “based on their own terms, and in reality most of our
life is in between and we cycle between these two things”.
Ray also suggested that our students should master a skill and
then share it with everyone else. They should find people who
have succeeded in a certain area - and know it better than anyone
- should repeat it.
ENTREPRENEUR
From the industry perspective, Andrea claimed that “You own your
career path. It is your responsibility and not anyone else’s – it
doesn’t belong to your boss, partner or family”. So, as it is a broader
field, students are recommended to filter what they like doing with
as much importance as what they don´t, because in the end they
will find the perfect fit.
So don´t despair if you haven’t found what you like doing, and as
well don´t be afraid to change the work you’re doing - that choice
“doesn’t need to be set in stone”.
Industry and big company level is a constantly growing type of
work and you have to “cultivate knowledge, dare to try new things
and make connections and build a network of contacts to succeed.”
INDUSTRY
In terms of work/life balance, although it depends on your
personal attributes and stage of your life - academia is more
balanced as working hours can be managed - while industry
has more specific schedules and the entrepreneurial sphere
is more flexible, it depends on the work rate you choose to
have and the objectives you want to reach.
We also talked about CV's in different areas and the
importance of publications. The latter is really important
to assess communication skills and knowledge on a specific
topic. You should also do research on the position and
institution to which you are applying - to be prepared to sell
yourself to that position and to understand the challenge
you are confronting.
Koen de Bosschere, HiPEAC coordinator, concluded the
session by challenging everyone to “have a dream and to
pick a role model”, as that will guide you to set up your
career path. Young people should start early to contact high-
ranking people in order to understand the professional world
in which they are interested. “Know your own value, don’t
be afraid of the unknown, be flexible, act professionally and
show leadership – and volunteer a lot”. If you show these
characteristics, other people will help you achieve your
goals. Lastly, “be passionate about your work” because if you
love what you do, you will do it gladly and with great results.
We would like to thank Andrea, Frank and Ray for sharing
their valuable knowledge with us, and hope that we can
meet them next year!
For more career-related information you can follow the HiPEAC
Jobs Twitter page at @hipeacjobs and check the multiple job and
internship opportunities at https://www.hipeac.net/
You can also watch the full session at: https://www.youtube.
com/watch?v=pQOGmX3p2sA&feature=youtu.be&t=1
“Master one thing better than your
peers, then share it with everyone.”
“Cultivate knowledge, dare to try
new things and make
connections and build a network
of contacts to succeed.”
15HiPEACINFO 61
Three-minute thesis
The HiPEAC network includes almost 1,000 PhD students who are researching the key topics of tomorrow’s computing systems. In this issue, we find out how heterogeneous parallel computing systems lend to the fast analysis of vast amounts of data
NAME: João Gante
RESEARCH CENTRE: INESC-ID/ Instituto
Superior Técnico, Universidade de Lisboa
SUPERVISORS: Leonel Sousa and
Gabriel Falcão
The Interplay Between Positioning and Beamforming in Millimetre Wave Communications
The contextOne of 5G's highlights is the introduction of millimetre wave
(mmWave) communications, unlocking untapped bandwidth.
However, with mmWave transmissions, the propagation
properties change dramatically: the resulting radiation not only
has severe path loss properties, but also reflects on most visible
obstacles. Fortunately, systems that work at those frequencies can
employ large arrays of antennas, on top of which beamforming
can be applied. Beamforming is a signal processing technique
that allows the system to steer and focus the transmitted beam
pattern, so as to mitigate the aforementioned drawbacks – for
instance, the radiation can be aimed towards a certain obstacle,
such that the resulting reflection reaches the desired destination.
The recent focus in mmWave communications also led to the
proposal of new positioning systems. The accuracy achievable
in certain conditions is remarkable, having sub-meter precision
in indoor and ultra-dense outdoor scenarios. However, there
was no mmWave positioning system that was able to perform in
outdoor scenarios under practical constraints.
Originality/ChallengesFrom past research, it was known that beamforming delivers
spatial selectivity. Our research started with the reverse question:
can we derive spatial information from beamformed signals? To
do so, we created a positional fingerprint signal, “beamformed
fingerprint”, which captures the power delay profile for multiple
beamforming patterns at a given position. If a dataset with the
sampled beamformed fingerprints alongside their corresponding
positions is known, then deep neural networks and supervised
learning can be applied. In fact, a significant part of the research
was with respect to the collection of said dataset, which implied
accurate 3D city models and precise propagation simulation.
Average error vs. average energy required per position fix for the
discussed outdoor positioning technologies.
Analogies/ApplicationThe concept of fingerprinting was known to the community,
and it is used to build low accuracy local positioning systems.
With beamforming, we can increase the spatial resolution
of the collected signal, akin to a LIDAR with higher angular
resolution. Likewise, the mmWave signal reflects on obstacles,
just like the light in LIDAR systems: we can see the proposed
work as an analogous system to collect information from the
surrounding environment, where the transmitter (base station)
and the receiver (mobile device) are placed in different locations
and the goal is to locate the receiver. In our work, the proposed
system achieves a smaller average error than low-power GPS
implementations, at 1.78 m. However, most importantly, it is
at least 47x more energy efficient than those implementations,
with larger gains for sporadic positioning uses, enabling smaller
devices to leverage positioning capabilities.
HiPEAC futures
HiPEACINFO 6116
Simulating COVID-19 and flu spread using HiDALGOOr how technology can support decision-making for an effective response to a pandemic
The current pandemic has highlighted the relevance of technology
for detection, prediction and prevention to support decision makers
in providing appropriate responses, taking into account health and
care capabilities.
In this sense, the European Union-funded project HiDALGO has
reused its expertise with agent-based modelling to develop FACS,
the Flu and Coronavirus Simulator, guided by the outcomes of SEIR
(Susceptible-Exposed-Infectious-Recovered) models operating at
the national level.
FACS approximates viral spread, incorporating geospatial data
sources from OpenStreetMap. In this way, the most crowded places,
such as residential areas, office buildings, parks or shopping centres
are identified as hot spots due to the high number of people who
can concentrate in them, as well as traceability routes to reach
these points.
So basically, COVID-19 spread is modelled at the local level,
providing estimations of the spread of infections and hospital
arrivals, given a range of public health interventions. The predictions
made by executing simulations can be used by decision makers to
identify peaks of contagion, set appropriate measures to reduce
spread and provide necessary means to hospitals to prevent
collapses.
The tool is publicly available under BSD 3-Clause license in GitHub
and the algorithms used to implement the simulations are explained
in this paper.
CPS Convergence and SocietyOn the wider Cyber-Physical Systems (CPS) stage in Europe, there
are some significant exchanges taking place in the final months of
2020 in preparation for the future of these products. Attention points
include orchestrating the research of this application domain,
evolving the supporting culture and measuring the CPS industry
pulse. Of particular focus is the adaptation of production and
assembly approaches, to be primed for new technologies, including
those required for restoring natural balance and minimising the
fallout from climate change.
A workshop was held in early September bringing together the two
communities of CPS and Systems of Systems (SoS). Prior consensus
from the CPS community indicated that one of the six key identifying
features of a CPS is the aspect of coordination and collaboration.
This means that SoS technology is a fundamental need for CPS with
implications for the way forward. There were two principle
discussions. One concerned the bottlenecks for the transfer of SoS
technology to CPS products and the other concerned the synergies
shared by the two communities – particularly related to the system
engineering challenges. A second consultation will take place in
November to support consensus before a formal report is published.
If you would like to participate, either check out the HiPEAC virtual
booth at EFECS for details, or contact us directly.
Back to the previously noted CPS focuses, an online HiPEAC
conference on CPS is organized by Tampere University in October.
There are plans for a session on ‘CPS and Sustainability’ in relation
to the Green Deal of the European Commission. We will consider
how the technology-pull landscape is evolving. This includes
interesting questions such as how CPS providers will need to adapt
for sustainability by increased long-term governance. There is also
the question of how CPS will need to better support sustainability in
society – for example wide-scale digital twin representation for
cause-effect analysis. Come and join us!
News
HiPEACINFO 61 17
The Computing Systems Week will be hosted
in Tampere, Finland. Due to the COVID-
19 situation, the event will be virtual. We
asked the local organization chair Prof.
Jari Nurmi of Tampere University what’s
up in Tampere.
What makes Tampere a good location for Computing Systems
Week (CSW)?
Tampere is the mother of industrialization in Finland, sometimes
called the “Manchester of Finland” because of its old red-brick
factory buildings. It has been the frontrunner in the development
and the adoption of many technologies, such as the introduction
of the first electric lights in Northern Europe and the famous
early “smartphone” (although they were not called smartphones
at the time) Nokia Communicator. Even the early days of Tampere
were very international; there were entrepreneurs, investors and
engineers from Germany, Russia, Sweden, Scotland, etc. Now the
old factories are mostly used as offices, restaurants, boutiques,
(cinema and live) theaters and museums. Information and
Communications Technology (ICT) is the industry of the day,
whereas the machine, paper and textile industries have mostly
disappeared. One big advantage of Tampere is its compact city
center; everything is within convenient walking distance.
What is the local technology ecosystem like in Tampere?
There are some big companies such as Nokia, but also a wide
ecosystem of smaller technology companies developing software,
hardware and services for the digitalizing society. The small and
medium-sized enterprise (SME) and start-up scenes flourish
on the ruins of the former Nokia mobile phone development
departments, ramped down after Microsoft acquired them. A
lot of engineering resources were suddenly available for new
companies to start and old ones to expand or transform their
operations in the city. There is also a long and strong tradition on
industry-academia collaboration in Tampere, especially involving
the engineering units of Tampere University (TAU).
What are some of the most interesting projects happening at
Tampere University?
There are the traditional strongholds of photonics and signal
processing (including Machine Learning). In the computing area
there are also the Transport-Triggered Architecture and Kactus2
open-source design tools, both with a long development history at
TAU. At the moment, there are multiple Marie Skłodowska-Curie
innovative training networks coordinated in Tampere, for instance
A-WEAR on wearable computing with security constraints,
ImmerSafe on immersive visual technologies for safety-critical
systems, and FibreNet on robotic systems for microscale material
testing. We will also start a new training network on Approximate
Computing in the autumn. There is also a unique concentration
of positioning research for both satellite-based navigation and
indoor localization. That is also reflected in the settling of many
positioning technology companies in the city.
What should we do and see in Tampere?
It is very unfortunate that due to the pandemic you cannot
try and see now all the typical things of Tampere, but luckily
there is a virtual replacement for that, Virtual Tampere. I hope,
though, that you will soon have another chance to visit the city
and get a real taste. You should try the traditional black sausage,
mustamakkara, though not everyone will be a fan! There are
also nice pubs with microbreweries in-house. There are a lot of
students in the city, so you might see groups of them in overalls
partying, any time of the year. In the winter season ice-hockey
is worth mentioning, there are two teams from Tampere playing
in the Finnish hockey league. What you would see without
even trying is the construction of two tram lines through the
city, scheduled to open in 2021, as well as a few other major
construction sites. There are some rare museums in Tampere, too,
such as the world’s first spy museum, an ice-hockey museum,
probably the only Lenin museum in the world, and the Moomin
museum. In the museum center Vapriikki you can see many
others, like the computer gaming museum and an exhibit on
the 1918 civil war, which came to its bloody end in Tampere.
Tampere is surrounded by two big lakes, and the rapids between
them cross the city center, providing nice sceneries for walking
by water. There is also the high sand ridge of Pyynikki with an
old observation tower on the top – most people go there for its
café with excellent doughnuts. Another place with great views
over the city and its surroundings is the Näsinneula tower, the
symbol of the city. From the bar at the top of the highest hotel
in Tampere you can also take a look over the whole city center. I
hope you can soon visit the city and experience these, all of them
are within walking distance!
Welcome to the virtual Computing Systems Week hosted by Tampere University
News
HiPEACINFO 6118
The ACACES summer school has run without interruption since
2005. This year, it had to be organized as a virtual summer
school. This change created challenges and opportunities: one of
the challenges was that the participants were no longer sharing
the same time zone, which created extra constraints, especially
for the teachers. We decided to organize a morning, afternoon
and evening track to accommodate teachers and attendees from
distant time zones. Another challenge was how to create the
summer school experience without being physically together.
That turned out to be much more difficult.
Fortunately, there were also opportunities. We could drop the
registration fee, which almost immediately tripled the number of
applications to the summer school. We could spread the summer
school over more days as we were no longer constrained by time
and space and we decided to run it as a single-track summer
school during two weeks instead of a three-track summer school
during one week. We now have recordings of every course that
we can use to further promote them.
After the summer school, we analysed the logs and we organized
a post-summer school survey. This led to some interesting
insights.
1. There were many no-shows, but we still had more active
participants than in the past. If we were to organize ACACES
again as a virtual summer school, we will consider charging a
low registration fee to reduce the number of no-shows.
2. The attendance in the first week was considerably higher than
in the second week. The extra week did not give us the extra
participation we were hoping for.
3. The appreciation score per course was very high and on
average higher than in the previous years. There is no reason
to assume that the quality of this year’s courses was different
from the quality in previous years. The explanation is probably
that participants only took (and evaluated) their favourite
courses while in a physical summer school, they take a full
program including courses that they would have skipped in a
virtual summer school.
4. Many participants mentioned that they could afford the virtual
summer school while they would never be able to afford the
physical summer school.
5. All recorded courses were published on the HiPEAC YouTube
channel after the summer school. Some courses have already
attracted more viewers on YouTube than there were participants
in the streaming version.
6. People who attended the summer school in the past very much
missed the social aspects of the school: living together for one
week, discussing research, and the food. We tried to encourage
interaction between participants with an instant messaging
platform, but that was no substitute for face-to-face discussion
at the summer school. Some students very much enjoyed the
scavenger hunt game organized by one of the teachers during
the break.
The conclusions are (i) that most attendees were very happy with
the virtual summer school, (ii) that the social dimension and the
networking was largely absent, (iii) that the recorded courses on
YouTube add value to the school. The virtual summer school was
a very interesting experiment, but in 2021 we hope to be able to
organize a physical summer school again.
https://www.hipeac.net/acaces/2020/#/
Afterthoughs on the ACACES2021 virtual summer school
News
HiPEACINFO 61 19