inside: your hipeac jobs poster
TRANSCRIPT
Inside:
your HiPEAC
Jobs poster
51APPEARS QUARTERLY | JULY 2017INFO
Team Europe: A workforce for the digital age
Martin Kersten’s forgetful databases
Made in Croatia: the world’s fastest electric car
Digital skills for a competitive Europe
Team Europe. Getting the brains to power the digital revolution
‘Data will rot away like everything in nature’
4 1512
contents
3 WelcomeKoen De Bosschere
4 Policy corner Digital skills for a competitive Europe
Heidi Cigan
12 HiPEAC voices ‘Data will rot away like everything in nature’
Martin Kersten
14 HiPEAC Jobs Trends in computing systems jobs
Xavier Salazar and Anna Molinet
15 Recruitment special Team Europe: Getting the brains to power the digital
revolutionSabri Pllana, Marisa Gil, Dionisios Pnevmatikatos, Alexey Cheptsov, Marc Gonzalez Vidal, Bev Bachmayer, Graham Mudd, Angela Bradfield, Meenakshi Ravindran and Karon Davis
22 Inside the box Well hello, AXIOM board!
Maurizio Caporali, Davide Catani and Xavier Martorell
24 Innovation Europe UpScale framework for real-time HPC applications
Luis Miguel Pinho
25 Innovation Europe InnoHPC brings HPC to the Danube region
Karina Pešatová
26 Innovation Europe The Heterogeneous Hardware and Software Alliance
Karim Djemame and Oliver Barreto
28 Peac performance Parallelizing Python codes using the superscalar
paradigmRosa Badia
30 Technology opinion Neuromorphic Computing: low-power systems,
the brainy wayKemal Delic and Dave Penkler
31 Technology entrepreneurship Keep the revolution going
Colin Adams
33 Industry focus Cache-aware Roofline Model in Intel® Advisor
Leonel Sousa, Aleksandar Ilić and Frederico Pratas
35 SME snapshot Rimac Automobili: The drive to succeed
Aco Momčilović, Matija Gracin and Marta Longin
36 HiPEAC futures Career talk: Michael Hübner, Ruhr-Universität Bochum HiPEAC collaboration grants: When Manchester met
‘Silicon Island’, Crete HiPEAC internships: Transport safety as standard Three-minute thesis: Memory access control which is
right on time
HiPEACINFO 512
The drive to succeedKeep the revolution goingParallelizing Python
353128
This issue is about digital skills. As the director of the computer engineering degree of
Ghent University for more than 10 years, this subject is close to my heart. Two large ICT
companies in Belgium have more open positions every year for computer engineers
than the total number of graduates in one year in Ghent. That is painful for them, but
they can still recruit internationally. Startup companies that do not find local talent face
more serious issues. Every month I receive emails from desperate startup founders who
have the money to hire computer engineers, but simply cannot find them. Hence, the
lack of qualitied ICT workers is slowing down innovation.
In 2016, Google published a report titled ‘Digitizing Belgium’, concluding that digiti
zation could lead to more than 300,000 jobs by 2020, or 100,000 per year from now on.
Yet we know how many computing experts will graduate in 2020 from the numbers
currently at university: at master’s level, there will be less than 1,000 for Belgium as a
whole. At best, taking into account all businessrelated ICT degrees and STEM degrees
that include decent software development skills, we might end up with 10 times more,
not 100 times more.
All wellintended digitization plans (the internet of things, Industrie 4.0, Smart Anything
Everywhere, etc.) will fail to deliver their full potential if we do not fix the workforce
issue first. Structurally increasing the number of graduates will take up to 10 years
because we have to start by sparking interest for computing at highschool level.
Retraining people could help in the short term, but a sixmonth training course will
never be a substitute for a four to fiveyear university degree. This is particularly
important given the complexity of modern information processing systems. Producing
inefficient and incorrect software leads to the accumulation of technical debt, to failing
systems, to security, safety and privacy issues.
Many of you will be reading this issue at the HiPEAC summer school, ACACES, which
is our contribution to training the crème de la crème of computing experts in Europe.
You will find a HiPEAC Jobs poster inside: by posting this at your organization, you will
also help ensure the best candidates find their ideal jobs.
We wish you a relaxing summer, and we hope to see you again after the summer
holiday in good health, and full of plans for the year to come.
Koen De Bosschere, HiPEAC coordinator
welcome
HiPEAC is the European network on high performance and 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. 687698.
Design: www.magelaan.be
Editor: Madeleine Gray
Email: [email protected]
HiPEACINFO 51 3
Digital technologies are advancing and
leading to the development of everything
from digital medical diagnosis and auto
nomous vehicles to drones and smart
fridges. Where digital was once a sector, it
is increasingly spreading to the whole
economy as society undergoes a digital
transformation.
This digital transformation of the economy
and society is also accelerating. The pace
of technological adoption is speeding up.
New technologies enter the market and
are being taken up more quickly than ever
before. For example, while it took decades
for the telephone to reach 50 % of house
holds, it took mobile phones less than five
years to reach the same penetration rate.
Today in Europe, there are 137 mobile
phone subscriptions for every 100 people.
Smart phones reached 40 % penetration
in just 10 years. New technologies such as
cloud services and social networks are
also being rapidly taken up. Indeed, in
2016, 62 % of internet users and 45 % of
enterprises in the European Union used
social networks.
To use these new technologies and ensure
individuals, companies, regions and coun
tries get the benefit from them, people
need to be able to use and work with
them. However, while technological
develop ment and adoption is speeding up,
people’s skills and the education and
training systems that are there to develop
them are slower to adapt.
Schools still lag behind in terms of digital
education. Equipment is still lacking in
many places and teachers, though often
willing, lack the skills and confidence to
make the most out of digital tools for
learning and teaching the next generation
how to use them. Digital tools are under
used in the classroom and most students
are still not taught the basics of computer
science and how to code – skills and com
petences that are increasingly becoming
essential for an understanding of the
world we live in. In higher education,
graduations in computer science have
fallen substantially over the past 10 years
and those students who do graduate are
often not considered work ready by
employers.
It is hardly surprising, then, that we do
not have enough of the skilled technology
experts – big data experts, cloud deve
lopers and cybersecurity experts – we
need for the new jobs emerging in the
economy. In fact, Europe already has a
large and growing deficit of digital experts.
From 373,000 in 2015, this gap could
grow to around half a million by 2020.
Digitally skilled workers are lacking at all
levels and in all sectors. For industry, and
the economy more broadly, to digitize,
workers in all sectors need to develop new
skills to work in the digitized workplace
and to remain employable. However, 37 %
of the labour force in the EU lacks basic
skills for the use of digital technologies.
Furthermore, while most companies are
aware that lacking digital skills in their
Digital skills for a competitive EuropeIn a special guest piece for HiPEACinfo, Heidi Cigan of the Directorate-General for Communications Networks, Content and Technology at the European Commission gives us the latest on the digital transformation, the need for new skills and the Digital Skills and Jobs Coalition.
Policy corner
HiPEACINFO 514
workforces negatively affect business per
formance, most do not provide oppor
tunities for their employees to reskill.
To support the development of digital
skills in Europe, on 1 December 2016 the
European Commission launched the
Digital Skills and Jobs Coalition. The
Coalition brings together stakeholders
from, for example, industry, education,
government and social partners to share
best practices and make ‘pledges’ to
provide training and carry out other
activities, such as awarenessraising activi
ties, to increase the digital skills of
European citizens.
Member States have been asked to develop
comprehensive digital skills strategies and
to set up national digital skills coalitions
involving a broad range of stakeholders in
order to support the implementation of
these strategies. The Commission has set
up a Coalition secretariat to support the
setup of national coalitions – for which a
small amount of ‘seed funding’ has been
provided – and other activities.
The Commission is also working with
Member States, through an expert group,
to support the development of national
digital skills strategies. The group has put
together a joint strategy document which
collects the main challenges faced and
solutions found by Member States in the
area of digital skills. A selection of best
practices has been linked to it and
published online as a tool to be used for
the development of digital skills strategies
and replication of solutions that work.
Since its launch, the Coalition has attrac ted
over 200 member organisations and around
70 pledges. Sixteen National Coalitions are
in operation or being developed.
The aim of the Coalition is to train over a
million people over the next two years, to
ensure Europe has a pool of digital experts,
and that workers and citizens more broadly
have the skills they need to succeed in the
digital transformation.
As a concrete contribution to this goal, on
23 March 2017, as part of the Digital Day
in Rome, the Commission announced the
launch of a new multimillion euro ‘Digital
Opportunity’ pilot scheme to provide
intern ships for current and recent
graduates of all disciplines in digital
domains. It is envisaged that these intern
ships should focus on ‘deep tech’ skills
such as such as cybersecurity, big data,
quantum or artificial intelligence, as well
as on more horizontal activities such as
web design, digital marketing, software
development, coding or graphic design.
The pilot scheme will initially run for two
years from 20182020 and will support up
to 6,000 paid internships.
To achieve our goals, we need to get all
stakeholders on board, and the HiPEAC
community can play an important role. You
are a large group of skilled professionals
with a wealth of knowledge in all areas of
digital technology, and you can potentially
make a huge contribution to the dissemi
nation of digital skills across Europe. For
many of you teaching is a primary activity,
so you know the challen ges ahead. We
invite you to join the Coalition and make
your own contribution to securing our
digital future.
MORE INFORMATION:
https://ec.europa.eu/digital-single-market/en/
digital-skills-jobs-coalition
https://ec.europa.eu/digital-single-market/en/
european-digital-progress-report
http://bit.ly/EC_Digital_Skills_Report
https://ec.europa.eu/digital-single-market/en/
digital-scoreboard
Digital skills for a competitive EuropePolicy corner
HiPEACINFO 51 5
HiPEAC news
HiPEAC on the road
Over the last few months, HiPEAC has been at
a number of key events in the computing
calen dar, spreading the word about HiPEAC
services and research by the HiPEAC community.
At Embedded World in March, HiPEAC hosted
a session on high-performance and embedded
architectures, and at DATE 2017 (Design,
Automation and Test in Europe) HiPEAC
organized a special session focusing on Euro-
pean research. Meanwhile, HiPEAC’s exhibition
booths at DATE and ISC High Performance
2017 featured presentations by European
projects and the HiPEAC mobile careers unit,
displaying computing systems vacancies from
across Europe.
HiPEAC also hosted a booth at the Digital
Innovation Forum in May, providing the oppor-
tunity to speak to business leaders about how
HiPEAC’s members can help digitize European
industry.
The HiPEAC Jobs mobile careers unit at DATE
2017
A warm Croatian welcome awaited Com-
puting Systems Week in Zagreb on 27 and
28 April, which brought together over 130
participants from academia and industry.
The event featured keynote talks on how
Rimac Automobili went from a hobby
project to building the world’s fastest
electric car and from Martin Kersten on
database management systems.
In his policy briefing, the European
Commission’s Panagiotis Tsarchopoulos set
out the direction of public HPC research
over the next few years. Advances in HPC
spurred by the European projects MANGO,
ExaNeSt, EcoScale and ExaNoDe were
presented, while EuroLab-4-HPC presented
HPC road mapping work and chaired a panel
session on open-source software for data
centres and HPC.
The event also featured a session focusing
on digital health – a key topic for the
HiPEAC community, given that health is the
second biggest market for embedded
systems in Europe. Representatives from
Ericsson Nikola Tesla, Xsensio and Bluebee
identified challenges and highlighted the
latest innovations in this field.
This Computing Systems Week also sought
to help early career computer scientists
develop their skills and explore career oppor-
tunities. With career advice and com pany
pitches, the ‘Inspiring futures! Researchers’
careers’ session offered students the chance
to find out more about different aspects of
their future in computing.
Meanwhile, the second edition of the
Student Heterogeneous Programming
Challenge gave students free reign to find
the best solutions for k-means clustering, a
popular approach for big data analytics and
machine learning. Marcos Canales, of the
'Heterogeniuses' team, commen ted that the
challenge provided 'a unique opportunity to
work with different people from the super-
computing area’, resulting in an experience
which was ‘truly enriching and represents
the spark of what may become future colla-
boration'. He added: 'The HiPEAC Student
Heterogeneous Programming Challenge is a
must for any student starting a career in
supercomputing.'
With the aim of helping researchers get
their message across, HiPEAC also organized
a media skills workshop at the event.
Learning from good (and bad) examples,
participants learned the basics of working
with the media, drafted press releases, and
gave mock interviews for radio and
television.
Thanks to local host Mario Kovac and his
team at the University of Zagreb for their
hospitality in the stunning Croatian capital.
See our interview with Martin Kersten on
p.12. For more on Rimac Automobili, see
p.25.
Hvala Zagreb!Digital health, the future of HPC and preparing new generations of computer scientists
Phot
o: E
nek
o Il
larr
amen
di
HiPEACINFO 516
HiPEAC news
European cooperation on connected mobility and digital industryDigital Day, which formed part of the 60th anniversary celebrations
of the Treaties of Rome, also saw 29 European countries sign a Letter
of Intent to intensify cooperation on testing of automated road
transport in cross-border test sites. The initiative aims to drive for-
ward plans in the European Data Economy strategy.
National elected representatives and industry leaders also pledged
to collaborate in the European platform on digitizing industry. The
platform will trigger collaboration and joint investments, provide a
forum to explore common approaches to regulatory problems and
help industries and countries exchange the means for workforce
re-skilling.
The event also saw the launch of the European Interoperability
Framework (EIF), which gives specific guidance on setting up inter-
operable digital public services.
Digital Day website: http://bit.ly/DigitalDay_EU
EU states to work together for world-class HPCOn 23 March, during Digital Day in Rome, ministers from seven
European countries (France, Germany, Italy, Luxembourg, the Nether-
lands, Portugal and Spain) signed a declaration to support the next
generation of computing and data infrastructures. The states will
work towards the establishment of a multi-government cooperation
framework for acquiring and deploying an integrated supercomputing
infrastructure for the exascale computing generation. Named
EuroHPC, the European Commission says that this project will be the
size of Airbus in the 1990s and of Galileo in the 2000s. It will provide
a major role for leading supercomputing centres and scientists in
Europe. In June, Belgium also joined the agreement.
HiPEAC co-founder and coordinator of HiPEAC 1, Mateo Valero, who
is the director of Barcelona Supercomputing Center, represented the
view from the supercomputing sector in a panel discussion on ‘How
Europe can regain leadership in HPC’. In his talk, Valero stressed the
importance of investing in hardware design in Europe to complement
existing strengths in software, mentioning how HiPEAC performs a
key role in promoting collaboration across the computing systems
spectrum. ‘If we don’t produce our own hardware, we won’t have
security […] and we also need to protect our companies,’ he
commented. For several years, Valero has been calling for the
establishment of an ‘Airbus-type consortium’ for HPC and big data,
in order to develop a full domestic exascale stack in Europe.
With a view to creating the chips of the future here in Europe, the
Commission included a challenge to develop a working prototype of
a low-power processor design in the ICT-05-2017 call for project
proposals (as reported in HiPEACinfo no.48). Successful projects will
be announced in the autumn.
Further information and a video of the panel discussion are
available on the European Commission website:
http://bit.ly/Digital-Day_HPC. See http://bit.ly/EuroHPC
for more details about EuroHPC.
Imec launches self-learning chip The Belgian innovation hub imec has demonstrated a self-learning neu-
romorphic chip. The chip, which imec says is a world first, is based on
OxRAM technology and has been shown to be able to compose music.
By co-optimizing the hardware and software, the chip concentrates
machine-learning characteristics in a small area, while consuming very
little power.
Ultimately, imec aims to achieve low-power, high-performance, intelli-
gent chips for a range of applications.
Visit the imec website for further information: http://bit.ly/imec_chip
HiPEACINFO 51 7
HiPEAC news
Rainer Leupers, RWTH Aachen University,
TETRAMAX project coordinator
Many HiPEACinfo readers will have heard
about, or even participated in, the TETRACOM
FP7 Coordination Action (www.tetracom.eu)
during 2013-2016. Its focus was on small to
medium scale academia-to-industry technology transfer projects (TTPs),
each one revolving around industrial uptake of specific software or
hardware intellectual property (IP) generated in research projects.
TETRACOM coordinated and sponsored 50 TTPs in total and had a
tangible impact on the European ICT landscape, as indicated by the
numbers given below.
We are excited to announce that this success story will be continued at
a much larger scale and with even higher ambitions during 2017-2021
thanks to the TETRAMAX project (www.tetramax.eu). Enabled by a total
budget of €7 million and 22 partners from all over Europe, this new
Horizon2020 Innovation Action will provide an implementation of the
European Smart Anything Everywhere (SAE) initiative in the domain of
customized and low-energy computing for cyber-physical systems and
the internet of things.
The project builds on three major activity lines:
1. Stimulating, organizing, co-funding, and evaluating different types of
cross-border technology transfer experiments, providing ‘EU added
value’ via innovative technologies to first-time users and broad
markets in European ICT-related industries.
2. Building and leveraging a new European competence centre network,
offering technology brokerage, one-stop shop assistance and training
in customized/low-power computing to small and medium-sized
enterprises (SMEs) and mid-caps. This will have a clear evolution
path towards new regional digital innovation hubs where needed.
3. Paving the way towards self-sustainability based on pragmatic and
customized long-term business plans.
The project impact will be measured based on well-defined, goal-
oriented performance indicators. The immediate aim is to support 50+
industry clients and third parties across the entire European Union with
innovative technologies. This will lead to an estimated revenue increase
of € 25 million based on 50+ new or improved customized, low-power
com puting products, the creation of at least 10 entirely new businesses/
SMEs, over 30 new permanent jobs and significant cost and energy
savings in product manufacturing. Moreover, in the long term,
TETRAMAX will be the trailblazer towards a reinforced, profitable, and
sustainable ecosystem infrastructure, providing competence, services
and a continuous innovation stream in customized/low-power com-
puting at European scale, yet with strong regional presence, as preferred
by SMEs.
Just like TETRACOM, TETRAMAX will work closely with HiPEAC via joint
event organization and dissemination. We also expect that the HiPEAC
community will be particularly interested in participating in our open
calls for co-funded technology transfer experiments.
Stay tuned for further announcements and pass this news on to your
European technology transfer partners, so as to be ready for the first
open calls to be published soon after the project kickoff in September.
www.tetramax.eu
TETRACOM evolves into TETRAMAX
TETRACOM’s vital statistics in tech transfer
HiPEACINFO 518
HiPEAC news
Turbo-charge your code with POP’s webinar series
Want to maximize your code’s performance?
POP, the EU-funded Centre of Excellence in
Performance Optimisation and Productivity, is
offering a series of webinars focusing on how
to optimize parallel codes for research and
industry. The webinars provide expert guid-
ance on code optimization, with practical
demonstrations and implementations on real
applications.
Coordinated by Barcelona Supercomputing
Center, POP brings together expertise from
across Europe. It offers services to provide
greater awareness of issues affecting comput-
ing performance, help manage the complexity
of modern computing and maximize the effi-
ciency of compute-intensive applications.
For further information, visit the POP website:
https://pop-coe.eu
New TaPaSCo tool composes accelerators into bitstreams for FPGAsJens Korinth, Technische Universität Darmstadt
The Embedded Systems and Applications Group (ESA) at TU Darmstadt is pleased to announce
the public release of TaPaSCo (Task Parallel System Composer), the successor to ThreadPool-
Composer, an open-source toolchain for creating task-parallel system-on-chip architectures
for hardware accelerators on FPGAs. TaPaSCo composes accelerators formulated in Verilog/
VHDL, or by high-level synthesis (HLS) in C/C++, into ready-to-use bitstreams for a range of
Xilinx FPGAs.
The tool can automatically scale the number of accelerator instances, aiming for the ideal
trade-off between area utilization and operating frequency. On the software side, the accel-
erators in the bitstreams generated by TaPaSCo are accessible using a universal easy-to-use
C/C++ API across multiple FPGA platforms. TaPaSCo currently supports PyNQ and ZedBoard,
Xilinx ZC706 Zynq and VC709 FPGA platforms, thus spanning the embedded, desktop, and
high-performance computing domains.
The tool is open source and has been designed with easy extensibility and modifiability in
mind. It is the result of three years of development in the EU FP7 project REPARA and is cur-
rently being used by a number of academic and industrial partners outside of TU Darmstadt.
The tool has proven especially useful in conjunction with HLS, since it easily closes the gap
between the hardware of individual accelerators and their integration into a hardware/soft-
ware system.
With the 2017.1 release of TaPaSCo, we cordially invite the research community to make use
of and participate in this project. Both the tool as well as introductory materials are available
from a dedicated GitLab site. We highly appreciate feedback, questions and contributions -
anyone is invited to contribute to the toolchain, be it extensions, modifications, bugfixes,
benchmarks, or other results. A bugtracker is available to request new features, or report
bugs; also, do not hesitate to contact us when you run into problems. Judging from the feed-
back of our existing users, TaPaSCo has significantly reduced their engineering efforts
required to make reconfigurable computing actually usable and avoided the need to reinvent
the wheel for the different computing platforms.
REPARA project: http://repara-project.eu
TaPaSCo GitLab site: http://bit.ly/TaPaSCo_GitLab
“POP webinars provide expert guidance on code optimization”
HiPEACINFO 51 9
Congratulations to Silexica, the RWTH Aachen
University spin-off specializing in multicore
software design automation, on being named
one of three finalists for the German Entrepre-
neur Award 2017 in the StartUp category. At
the time of going to press the winners had
not been announced.
In April, the company released the latest ver-
sion (2017.4) of its SLX Tool Suite. This
release provides many new features to
improve the efficiency of multicore program-
ming and code distribution. These include:
• The SLX Parallelizer now has a powerful
reconfigurable cache analyser that simu-
lates cache behaviour and estimates usage
statistics.
• The SLX Mapper has a new set of visuali-
zation capabilities providing deeper
insights into application runtime behav-
iour of computed/simulated mappings.
• The SLX Generator now features added
support for additional processor cores
such as 32-bit ARMv7 and 64-bit ARMv8,
as well as Linux-based Power-PC 32-bit
and 64-bit processors.
• The SLX Automotive Development pack-
age is now available with the rest of the
SLX tools.
silexica.com
Find out who was named German Entrepre-
neur 2017 on the awards website:
www.deutscher-gruenderpreis.de/en
HiPEAC news
What the F-OMP is going on in OpenMP?F-OMP: A Feedback monitoring infrastruc-
ture for OpenMP on embedded systems
Giacomo Valente, University of L'Aquila
Embedded systems normally execute appli-
cations with both functional and non-func-
tional constraints, and the underlying hard-
ware and software can be very complex and
heterogeneous. Different techniques are
used to manage the complexity of develop-
ing a system which meets the necessary
constraints. Among them is the use of
OpenMP, which supports designers in paral-
lelizing applications written in C/C++ and
Fortran code.
As OpenMP allows implicit parallelism, it is
harder to control all the factors involved in
the performance, such as memory accesses,
cache behaviour and thread mutual exclu-
sions. To solve this, F-OMP proposes a moni-
toring infrastructure which provides feed-
back about the use of OpenMP in an
embedded platform, developed on FPGA.
Specifically, it uses metrics to organize data
collected at runtime into useful information
(estimated speed-up, load balancing and
false sharing), providing feedback to the
designer without inserting software overhead.
We have implemented F-OMP on a Zynq7000
SoC, with a dual-core ARM processor as the
master processing element and four-core
Leon3 SMP as an isle of computational ele-
ments. This isle executes a Linux operating
system, which provides support for OpenMP
applications, and it executes two OpenMP
benchmarks with data provided by the
master processing element. The F-OMP
system was inserted in the platform, provid-
ing the metrics listed above.
Silexica’s Maximilian Odendahl and Johannes Emigholz
Silexica reach German Entrepreneur Award finals
HiPEACINFO 5110
HiPEAC news
Roger Needham Award presented to Alastair Donaldson
Congratulations to HiPEAC member
Alastair Donaldson, who has been pre-
sented the BCS Academy of Computing
Roger Needham Award 2017. The award,
which is sponsored by Microsoft Research
Cambridge, was given to Donaldson ‘in rec-
ognition of his outstanding work in the area
of many-core programming’.
According to BCS, Dr Donaldson, from the
Department of Computing at Imperial Col-
lege London, has ‘made a distinguished
contribution through his design and appli-
cation of rigorous program analysis meth-
ods to the emerging field of many-core
programming. His techniques and case
studies have also made a major contribu-
tion to fundamental Computer Science
research’. Dr Donaldson’s work centres
around formal verification, systematic test-
ing, programming language design and
compiler technology.
On receiving the award, Dr Donaldson said:
‘I'm incredibly honoured to be receiving
the 2017 Roger Needham Award. It's really
exciting to have my work recognized via
the award, and more generally the award
emphasizes the importance and potential
of research in programming languages and
analysis and verification.’
Further information can be found on the
BCS website: http://bit.ly/
RogerNeedham_AlastairDonaldson
Amir Ashouri wins Italian IEEE Computer Society PhD Thesis Award
HiPEAC student Amir Ashouri’s PhD thesis has been
selected as the winner of the 2016 IEEE Computer
Society Italy Section Chapter PhD Thesis Award. His
winning thesis, titled 'Compiler Autotuning Using
Machine Learning Techniques', was written under
the supervision of professors Cristina Silvano and
Gianluca Palermo at the Politecnico di Milano, as
well as John Cavazos of the University of Delaware.
Congratulations to Amir on winning this award!
Further information can be found on the IEEE Computer Society website:
http://bit.ly/IEEE_CS_Italy_PhD
Mateo Valero wins IEEE-CS Charles Babbage AwardMateo Valero, professor in the Computer Archi-
tecture Department at the Universitat Poli-
tècnica de Catalunya – Barcelona Tech and
director of Barcelona Supercomputing Center,
has been awarded the 2017 IEEE Computer
Society Charles Babbage Award. According to
the IEEE, this award recognizes Valero’s ‘contri-
butions to parallel computation through brilliant
technical work, mentoring PhD students, and
building an incredibly productive European
research environment’. The award was presented
at the annual IEEE-CS Inter na tional Parallel and
Distributed Processing Symposium (IPDPS
2017), where Valero presen ted a keynote
speech on runtime-aware architectures.
Valero, who co-founded HiPEAC and coordi-
nated the first phase of the project, has won
numerous awards over his career. These
include the 2007 IEEE/ACM Eckert-Mauchly
Award, the 2015 IEEE-CS Seymour Cray Award,
the 2009 IEEE Harry Goode Award, the 2012
ACM Distinguished Service Award and the
2015 Euro-Par Achievement Award, in addition
to multiple awards in his home country of
Spain. He has been given honorary doctorates
by nine universities and was admitted to the
European Union’s ICT (Innovate, Connect,
Transform) Hall of Fame in 2008 as one of the
25 most influential information technology
researchers in Europe. He is a member of five
academies and the school in his hometown,
Alfamén, has been named after him.
Congratulations from the HiPEAC community
on this award!
HiPEACINFO 51 11
‘Data will rot away like everything in nature’
HiPEAC news
What does Monet DB offer that other
database technologies don’t?
Monet DB was developed over a period of
twenty years, and its origin was the
decision to change the storage format
from a recordoriented mechanism to a
column structure. When we started the
project in the 90s, this was a nogo area
proven by science. It took us seven years
to prove that it was in fact possible. An
enormous field opened up called data
analytics, where this was the perfect
solution. Since around 2010, all major
database founders have incorporated
columnoriented technology, which we
pioneered, into their product lines.
What is your favourite application of
MonetDB?
Actually, it’s the product itself. I’m really
proud that we were able to build Monet
DB: we worked for it for about 20 years
and I’m really pleased that for all those
years I had a team of very strong hackers
dedicated to making it a viable product.
What’s the most challenging database
project you’ve ever had to work on?
A milestone was in 2005, when we
thought we had a finished product which
we could ship to the outside world. We
found an example in a big catalogue of
astronomical data that was developed by
Jim Gray and Alex Szalay. It was a pivotal
example because it showed that astro
Is it time to rethink our approach to data storage? In this interview, Martin Kersten (Centrum Wiskunde & Informatica) gives us the lowdown on Monet DB, how to deliver the memory for exascale computing and why we should think about whether we really need to keep data forever.
The astronomy field generates huge amounts of dataPhoto: ESA/ATG medialab; background: ESO/S. Brunier
HiPEACINFO 5112
‘Data will rot away like everything in nature’
HiPEAC news
nomers could actually do their research on
a database. As a database was available in
public, we thought, ‘if Microsoft SQL
server can do it, so can we’. So we
reimplemented the existing application
using our technology.
In the end, we were the only ones who
successfully mimicked that approach, and
there were some major hurdles along the
way; but the outcome was that all the
technology was available in Monet DB,
open source, even if you did have to pay
Microsoft for the SQL server licences.
Over the last five years, Monet DB has
opened up the whole software ecosystem.
The database system is not a black box,
where you have to use poor APIs to store
and retrieve data. Instead, we have
achieved the blending of your favourite
programming languages, your favourite
libraries, inside the kernel. This signi fi
cantly improves the effectiveness and the
performance of your application/database
combination.
Why is column-store important?
Going from a roworiented representation
to a columnoriented representation, you
get leaner files of data, which gives you
many more opportunities to compress
your data. Compression in a column store
is often better than you would see in
traditional compression structures. This is
relevant because if you can make things
small they require less storage and less
communication to get the data from the
disk into the memory for processing.
The second aspect of the system is that the
execution method used in Monet DB is
completely different from other database
systems. It is much more attainable for
parallel processing. For the details, see the
literature.
How is ExaNeSt (European Exascale
System Interconnect and Storage)
paving the way to exascale computers?
Exascale storage is already there; in fact,
my own lab alone probably has a couple of
petabytes of storage. In ExaNeSt, our
vision is that we can have hundreds of
thousands of processors with limited
memory. This allows us to give people a
small slice of that machinery to do
database processing without any concern
about others trying to use the same
machinery, meaning that you don’t have
competition for resources.
The underlying technology for that is
elasticity of the system, so that it is easy to
repartition your data, your storage plat
form and your compute platform for
particular uses. This is not something you
do every week, or every night – you may
have to do it every minute while your sys
tem is running. So ExaNeSt is about pro
viding extreme elasticity, with a focus on
providing the user the amount of resources
they need to do their job – quickly.
What’s the biggest challenge in the data-
base world over the next 10 years?
There are many; we’ve published a couple
of vision papers on the topic. The latest one
I put on the table for the database commu
nity itself is to stop keeping data around
forever and ever. Rather, we should start
out from the other end of the spectrum and
say, ‘data will actually lose its value, it will
rot away like everything in nature will rot
away’. So we should design a system using
a fundamentally different mechanism.
What this will lead to is that we ask our
users to be more concerned about how
long the data should be maintained, or to
distil it into summarized information. A
good example is to go back to the
astronomers: in the astronomy world,
they collect years of the light intensity of
stars. You can store that – but every star
can actually be approximated with a
Gaussian distribution. This can be stored
in just a few bytes.
Now, wouldn’t it be nice if we took that
enormous catalogue of observations and
turned it into a collection of Gaussian
models? This actually requires less than
one per cent of the storage, but an
astronomer could no longer distinguish
whether they were looking at the actual
data or at models of the data. That would
be my dream for the future: can we
somehow compress the data into much
smaller, manageable and more valuable
pieces of information. In fact, as
individuals we don’t want our personal
data to be retained forever, so having a
database system which starts forgetting by
itself would be a bright future.
www.monetdb.org
www.exanest.eu
HiPEACINFO 51 13
02005
100
200
300
400
500
600
700
800
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
9 4 14 17 15 2365
134176
198
351
525
600
PhD Student
OtherMaster
SoftwareEngineer
Engineer
CompanyResearcher
ResearchAssociate Assistant
ResearcherSenior
Researcher
LecturerAssistantProfessor
AssociateProfessor
Professor
PostDoc
C++ Python
SYCL
Javascript
Verilog
Caffe
OmpSs
MPIVHDLOpenCL
JAVA
CUDA
Hadoop
OpenGL
OpenMPLLVM
Tensorflow
Number of vacancies
Levels of experience
Vacancy locations
IoT (77)CPS (33)
Bioinformatics (31)
Fintech (1)
Machine Learning (114)Deep Learning (28)
Artificial Intelligence (16)
Energy (378)
Smart Cities (6)
Computer Vision (50)
Healthcare (46)
Big Data (93)Data Analytics (46)
Automotive (75)
Climate (52)
Timing Analysis (14)
Avionics (16)
Space (144)
Cloud (229)
Technical skills Technological Expertise
Application areas
HiPEAC jobs
Trends in computing systems jobsWith 10, 000 visitors every month, a number that continues to grow,
the HiPEAC Jobs portal is an indispensable resource for anyone looking
for a new opportunity or seeking high-quality staff for specialist com-
puting roles. Over the years it's become a key tool to take European
research forward: in total, 411 jobs related to European projects have
been published on the portal. The vacancies are published on the web-
site for an average of 65 days, which indicates that the technical skills
in demand may be in short supply and provides an indication of the
time needed to find the right candidate. To help spread the word about
the portal, we invite you to display the HiPEAC Jobs poster in this issue
on information boards at universities and schools.
Numbers provided indicate total number of vacancies posted on
the portal
hipeac.net/jobs
450
1
HiPEACINFO 5114
Recruitment special
As Heidi Cigan indicates in her article (pp.45), the digital skills
gap is approaching crisis levels in Europe. In A New Skills Agenda for Europe (2016), the European Commission estimates that
unfilled vacancies for ICT professionals could almost double to
756,000 by 2020. Yet, as the HiPEAC Vision 2017 notes, from
20072013 the number of European graduates in computing only
grew by about 0.5% per year, and even declined in several major
European states.
Europe boasts 42 of the world’s top 100 universities. Although
the quality of education provision varies across the region and
education providers struggle to keep up with the rapid pace of
technological change, many have access to a highquality
Team EuropeGetting the brains to power the digital revolution
education here. However, many skilled graduates leave for
opportunities outside of Europe, perhaps due to the lack of a
digital ecosystem, with major ICT companies being located
outside of Europe. An absence of highprofile industry
spokespeople may also contribute to computer science being less
attractive as a career choice in Europe.
In this article, HiPEAC experts give their opinion on everything
from getting children interested in computer science, to university
curricula, to the unique advantages of working at a smaller
company or a global brand. We also explore European initiatives
promoting excellence in human resources and discuss how to get
more women on board.
A CHALLENGE FOR THE FUTURE
Sabri Pllana, Associate Professor, Department of Computer
Science, Linnaeus University
Europe is already doing well in terms of
computer science education, but we lack
prominent role models. We need to do
more to promote entrepreneurs who have
studied computer science in Europe and
achieved global success in ICT through
the mainstream media.
To attract children’s interest in computer science, it is important
to establish a link between theoretical knowledge and its
application in solving realworld problems early in the learning
experience. We believe that appropriate pedagogical methods
can increase children’s motivation. For instance, competitions
can promote computer science and discover talent among young
people.
SciChallenge aims to boost children’s interest in education and
careers in science, technology, engineering and mathematics
SciChallenge is inspiring young learners in STEM subjects Photo on top © Kinderbüro Universität Wien (Barbara Mair)
HiPEACINFO 51 15
Recruitment special
(STEM) through a panEuropean contest. Contestants were
asked to submit a poster, presentation or video on any STEM
related topic, although SciChallenge suggested 50 topics as
inspiration. Many of these are related to computer science,
including the internet, robotics, cryptography, 3D printing, or
ehealth.
The contest is designed with gender equality in mind and,
interestingly, more than 50% of the contestants are female. One
presentation created by a group of girls for SciChallenge has
attracted over a million views (http://bit.ly/SciChallenge_
ImmortalGirls). The success of female contestants in SciChallenge
will inspire other girls in future.
The results have been inspirational, and there also some potential
future HiPEAC members among the candidates, such as:
• the creators of a device to transport industrial components
using sensors, vacuum grippers and a programmable logic
controller: http://bit.ly/SciChallenge_SLO_SRB
• a group which used machine learning and data analytics to
predict a musician’s future popularity: http://bit.ly/
SciChallenge_Huemer_Jetzinger
• a contestant who developed an Arduinobased motorized
gimbal for camera stabilization: http://bit.ly/
SciChallenge_TadejStrah
TRAINING THE NEXT GENERATION
Marisa Gil, Associate Professor,
Department of Computer Archi tecture,
Universitat Politècnica de Catalunya
– Barcelona Tech
We need a welltrained workforce in
highperformance computing (HPC) and
embedded systems here in Europe because the problems we face
and the continent’s particular characteristics are different to
those of the United States or Asia, for example. Having
homegrown technology professionals helps ensure that the
objectives, challenges and solutions are appropriate and are
those which will really lead Europe to advance.
As a network of computer scientists, within HiPEAC we have a
responsibility to train and nurture future generations, leaving
successors who will continue the work in this area. Educators
need to be capable of training students to be open minded and
able to identify the main issues and objectives. We should
Tadej Strah built a motorized gimbal to stabilize cameras
The full list of finalists is available on the SciChallenge website:
www.scichallenge.eu
SciChallenge is funded by the European Commission’s Horizon 2020 research and innovation programme under grant agreement no 665868.
encourage students to be aware that people and their needs
come before solving technical problems which won’t make the
world simpler and better for everyone. They should be agile and
unafraid of the effort required to effect real change and create
new ways of thinking.
To push their career forwards, students should first undertake a
thorough and personal examination of how they can make the
most of their studies and the opportunities available to them.
Next, they should share their experiences with students and
senior staff. Finally, they shouldn’t be afraid of being a bit short
of money or not knowing exactly what they will do in the future;
they will certainly achieve great things.
Marisa Gil organizes the Student Programming Challenge at Computing Systems Week with Chris Fensch of Heriot-Watt University and Georgios Goumas of the National Technical University of Athens (see p.6).
HiPEACINFO 5116
Recruitment special
CURRICULUM MATTERS: EUROLAB-4-HPC
Dionisios Pnevmatikatos, Affiliated Researcher, Foundation
for Research and Technology – Hellas (FORTH) and Alexey
Cheptsov, Research Scientist, High Performance Computing
Center Stuttgart (HLRS)
Since the introduction of cloud technology, highperformance
computing (HPC) is no longer limited to a small set of experts
but is visible to and affects even everyday users of desktop PCs
and mobile devices. In recent years, HPC has gone beyond
traditional application areas such as climate analysis, weather
prediction, astrophysics, etc., to new applications in social media,
big data, biomedicine and many more. Many of these new
applications are being pioneered by small and medium
enterprises, facilitated by European Commission initiatives such
as the FORTISSIMO project (www.fortissimoproject.eu).
Research, science and education are the main pillars to ensure
that leadingedge services are delivered in all areas of HPC
provisioning. While HPC is well positioned in terms of research
and science, the educational HPC horizon remains limited.
Unfortunately, there is currently no common view among leading
European educational institutions on the process of expanding
curricula to include HPCspecific topics. Standardi zation in HPC
curricula in Europe should start at the basic level, so that the
foundational set of ideas, terminology, skills, and tools are taught
to all students.
Most current higher education curricula do not stress parallel
thinking from the beginning and more advanced topics are
usually only included in specialized programmes. We need to
increase understanding of parallel programming and algorithms
in the general student population. A good starting point would
be to present parallel computing using students’ mobile phones,
which are usually built with manycore technologies, then
gradually moving out to the scale of modern supercomputers.
‘Think parallel’ should be the main message communicated to
students of computer science and related subjects.
To help address this issue, EuroLab4HPC has been working on
a model HPC curriculum. We have been collecting material from
existing educational programmes around Europe and beyond.
We have also used a questionnaire to find out the opinion of
experts and educators in the field, with their feedback being used
to better focus the direction of the proposed curriculum. In
addition, we collected information about existing courses at
institutions around Europe.
The major educational targets we have set for the curriculum
are:
• to promote parallel thinking
• to develop the ability to program parallel systems
• to promote understanding of performance (in processors,
compilers, code, communication, etc.)
• to evaluate and debug performance
• to promote knowledge of parallel algorithms
• to promote programming and use of heterogeneous systems
Our aim is for the proposed curriculum to act as a reference point
both for educators and institutions that would like to extend
their course inventory into HPC. We also hope it will allow
students to select courses, either within their institutions or via
distance learning, to build their knowledge and skills in parallel
and HPC programming.
HPC experts! Want to help shape the HPC curriculum? Complete the online survey here: http://bit.ly/2n5Pvy7
EuroLab-4-HPC is funded by the European Commission’s Horizon 2020 research and innovation programme under grant agreement no 671610.
HiPEACINFO 51 17
Recruitment special
Being a senior researcher at an institution like BSC offers the
opportunity to work on international projects in a unique
environment with toplevel facilities and interdisciplinary teams.
To progress to research management level, candidates need a
range of skills and experience.
In terms of research, an international reputation based on
research excellence in their field is fundamental, as is critical
judgement in the identification and execution of activities.
Researchers need to have made a substantial contribution
(breakthrough) in their field or across multiple areas. They will
have published and presented influential papers and books,
served on workshop and conference organizing committees and
delivered invited talks. They should recognize the broader
implications and applications of their research and develop a
strategic vision on the future of their research field.
However, the skills to progress go beyond the research itself.
Senior researchers are increasingly required to secure funding
and to manage and lead research projects. As team leaders, they
will need skills in managing and developing others. In addition
to team building, they will need to focus on longterm team
planning, such as career positions for researchers and securing
funding for team positions. Creative and innovative, they will
need to be excellent communicators and networkers in the
research community and beyond, and to act as professional role
models.
At BSC we use a tool to help identify each individual’s develop
ment needs and use the information gained via this tool to
shape our development programme. Researchers, with the
support of their managers, can choose from an extensive
programme of training and development actions including
project management, leadership, scientific writing and time
management, among others.
DEVELOPING THE SKILLS FOR SUCCESS
Marc Gonzalez Vidal, Head of Human Resources, Barcelona Supercomputing Center (BSC)
HiPEACINFO 5118
Recruitment special
POWERING A GLOBAL ELECTRONICS LEADER
SERIOUS FUN AT CODEPLAY
Graham Mudd, Senior Technology Manager, and Angela Bradfield,
Head of Human Resources, Samsung Research UK (SRUK)
The more niche and specific a skillset is for a particular role, the
more challenging it can be to find candidates who meet the
particular requirements. Meeting talented engineers is therefore
one of our highest priorities and greatest challenges: we know
the right people are out there and we cast our net far and wide to
find them. Another challenge we face is communicating exter
nally the fact that, while we benefit from the strength of the
global Samsung brand, our UK R&D site is just over 200 strong,
so we have the advantage of a boutique, familyfeel environment.
It’s a bit of a cliché, but we really look for talented people who
love solving problems and having their solutions show up in
actual products. We often hear employees describe the pride they
felt when purchasing a Samsung device from our staff shop for a
family member or friend and being able to brag that ‘I helped
build that’. Our core domains are embedded graphics, computer
vision, augmented reality, machine learning, digital TV and the
internet of things. At SRUK, you could be working on projects
such as machine learning for power management, stabilizing
virtual reality video or creating a lowlevel graphics API, for
example. We’re a hugely diverse bunch, with over 30 nationalities
represented in the SRUK office, so specific experience, personality
and professional formation are of less importance.
To get the best employees, the three watchwords are edu cation,
environment and enterprise, with commercial organizations,
policy makers and private individuals and associations all having
their part to play. Our strength lies mainly in the area of
enterprise, but we also need access to highquality technology
education for all and for there to be healthy, sustainable and
welcoming local environments in which to live and work.
Keeping people happy and engaged is one of the keys to
Samsung’s global success. Some of the ways we try to achieve
this at SRUK include good compensation, a great working
environment, continuous learning and development opportunities
– whether you decide to focus on the technical or management
route – and the ability to work on gamechanging projects that
will most likely show up in a multimillion unit device.
More detailed information about roles at Samsung can be found on the Samsung Careers page http://bit.ly/SamsungUK_Careers
Meenakshi Ravindran, Staff Software Engineer, Compilers,
and Karon Davis, VP Operations, Codeplay
We work on cuttingedge tech
nologies revolving around
hetero geneous com puting,
parallel computing, providing
a highperfor mance platform
for machine learning, artificial
intelli gence, computer vision
and other computationally intensive graphics applications. These
require technical skills including a strong core knowledge of
modern C++ (11/14), coupled with understanding of closeto
themetal software design, for example how hardware is
programmed at low level for optimized performance. That could
include areas such as compiler debuggers, programming
language design, memory management, graphics programming,
or parallel software design.
Operating in such a highly specialized field, it is quite challenging
to find candidates with the required skills. These skills are in
short supply so we have a global reach: we are very proud of the
international diversity of our team and encourage applications
from all over the world. We also have very good relationships
with universities and organizations like HiPEAC; our internship
programme is well respected, with internships available at both
undergraduate and postgraduate level.
Specific policy measures could boost the number of candidates.
These could include fast tracking visa applications for the people
we need to build our company, encouraging students to study
science and technology at school and supporting girls to succeed
in these areas. It might be worth introducing financial incentives
for students to study relevant subjects.
The single biggest advantage of working at a smaller company
like Codeplay is that, as an individual, your voice and opinion are
always considered and taken into account. You therefore have
the ability to contribute and truly feel part of the company rather
than getting lost in a corporate conglomerate. There are oppor
tunities to take on interesting projects and advance your career
by continuously increasing your knowledge and expertise.
To help settle new colleagues, Codeplay offers a threemonth
mentoring programme to all interns and new starters, whatever
their experience level. We have a robust knowledgesharing
HiPEACINFO 51 19
Recruitment special
The HR Strategy for Researchers at BSCBSC is the leading supercomputing
center in Spain, hosting MareNostrum,
one of the most powerful super
computers in Europe. With over 500
members of staff from 45 countries,
BSC combines HPC service provision
and research and develop ment (R&D)
activities under one roof.
BSC is committed to the principles of the EURAXESS European
Charter for Researchers and the Code of Conduct for the
Recruitment of Researchers. Via the HR Strategy for Researchers,
BSC aims to maintain the best environment for research
excellence, continuously improving through the adoption of
international best practices and highquality standards.
The HR Strategy for Researchers supports organizations to
implement the Charter and Code in their policies and practices.
Concretely implementing the Charter and Code makes research
institutions more attractive to candidates looking for a new
employer or a host for their research project. Funding organi
zations which implement the Charter and Code principles
contribute to the attractiveness of their national research
systems and of the European Research Area more generally. As
a result, the HR Excellence in Research logo identifies
organizations which either provide or support a stimulating
and favourable working environment.
The HR team at BSC
In April 2015, BSC was awarded the HR Excellence in Research
logo, recognizing BSC’s commitment to the principles of the
Charter and Code. This commits the centre to an action plan,
developed inhouse and approved by the European Commission,
to improve its capabilities and performance across a wide range
of training, recruitment and gender issues. The plan has since
been revised by the HRS4R steering group based on an analysis
of the results from 20152016, and a new plan for 20172020
has been created and approved.
Highlights of this ambitious project include:
• the creation of advisory committees
• boosting equal opportunities and gender policies
• strengthening activities in outreach and public engagement
• improving the BSC career development system
• incorporating the EURAXESS Open, Transparent and Merit
based Recruitment of Researchers (OTMR) practices
European Charter for Researchers https://euraxess.ec.europa.eu/jobs/charter
programme and strongly encourage all staff to participate in
local and international communities and events. We also organize
regular social events, from games over lunch and weekly movie
nights to annual company gatherings. We’re very proud of our
record in welcoming staff from over 18 different countries and
helping them to settle in quickly to their new country and role.
For the latest career opportunities, visit the Codeplay websitehttp://bit.ly/Codeplay_Careers.
Find Codeplay, SRUK and many more vacancies on the HiPEAC Jobs Portal: hipeac.net/jobs
HiPEACINFO 5120
Recruitment special
Finding Nema: breaching the ICT gender gapWe caught up with Bev Bachmayer, Vice Chair of ACM-Women Europe (ACM-W), to find out how we can help get more women in ICT.
What are some of the reasons behind
the lack of gender balance?
There are hundreds of studies attempting to explain this
phenomenon. People point out factors such as the difficulty in
achieving work/life balance, problems arising from the male
dominated work environment or blatant bias where companies
do not enforce gender equality regulations. However, it’s not
just the front of the pipeline but the leaky pipeline which finds
us in this predicament.
To combat the problem, we need to focus on the barriers. One
is unconscious bias, which is the perception that women are not
engineers or scientists. A second barrier is the lack of role
models: young impressionable women need to see people like
them doing the job, so that they can identify that they could do
that too. Lastly, a major problem is when qualified women end
up leaving the field because of a multitude of problems.
Why is it important to provide support specifically aimed at
women?
Networking helps women attain higher levels of achievement.
Women often feel isolated when they are the only female in a
group. Bringing technical women together gives them
opportunities to see that they are not alone. Additionally, seeing
successful women talk about issues that are part of many
women’s daily life inspires young women to continue in the
field. Feedback from womENcourage events shows that 89% of
the attendees are more committed to their careers after they
attend womENcourage.
Why should men support initiatives to increase the number
of women in computing systems jobs?
We need to change the culture and men are needed to help
make that change. Many people are unaware of some bias they
may hold, thereby making it difficult for another colleague to
excel. By joining the discussion and supporting such initiatives
people will see that the unconscious bias exists.
Is there a danger that such initiatives unfairly promote
women?
As soon as 50% of working ICT experts are women, conferences
have 50% women technical experts on the programme and
50% of computer science/engineering degrees are earned by
women, then we can worry about whether we are unfairly
promoting women.
What practical action can organizations take?
To recruit:
1. Participate in Celebrations of Women in Computing and
careers fairs (see below).
2. Bring your engineers to these events to network with the
participants.
3. Offer internships.
4. Provide mentors for students.
To retain:
1. Promote work/life balance, flexible working hours, working
from home.
2. Improve maternity leave.
3. Provide child daycare.
4. Proactively build the team with social activities that are
inclusive for all at work.
ACMW organizes regular Celebrations of Women in Computing.
In addition to keynote speeches, panel discussions, workshops
and hackathons, these events feature a careers fair where
supporters can recruit students and professionals. Supporters
this year include HP, Google, Bloomberg, Oracle, Accenture,
Intel, Amazon, Informatics Europe, Inria and Microsoft.
womENcourage 2017, 6-8 September, Barcelona, Spain https://womencourage.acm.org/
HiPEACINFO 51 21
After almost two years, we’re delighted to
present the AXIOM Board, Serial Number:
000000000001. It is the initial result of
thorough research funded by the European
Commission and involving seven different
entities across Europe: three research
groups/universities – the University of
Siena, Barcelona Supercomputing Center
and the Foundation for Research and
Techno logyHellas (FORTH) – and four
enter prises – SECO, Vimar, Evidence and
Herta Security.
This is the first version of the AXIOM
Board, but we truly think we’re on to
something. AXIOM is the first board that
combines three worlds in one: Arduino,
ARM computing and FPGA. It runs a
version of BSC’s flagship programming
model, OmpSs, allowing straightforward
FPGA programming. It’s designed to be
the perfect combination of high
performance computing, embedded
computing and cyberphysical systems. As
such, it aims to provide the ideal platform
for realtime data analysis of a huge
amount of data in a short time frame,
machine learning, neural networks, server
farms, bitcoin miners – you name it.
Take a look at the image on the opposite
page to see what marks the AXIOM board
out from the competition.
AXIOM is funded by the European Commission under the H2020 Framework Programme for Research and Innovation under grant agreement no. 645496.
Inside the box
Well hello, AXIOM board!
Last year (issue 47) we caught up with Maurizio Caporali (Università di Siena) and Xavier Martorell (Universitat Politècnica de Catalunya – Barcelona Tech) to find out about the AXIOM project’s plan to build a European single-board computer. We were very excited when the board finally arrived this year. Along with Davide Catani (SECO), Maurizio and Xavi offer us a peek under the hood to show why this board is so special.
The AXIOM board was debuted at Embedded World 2017
HiPEACINFO 5122
Innovation Europe
In this issue we find out about a design framework for applications with strict timing and high-performance requirements, a project to bring HPC to the Danube region and the Heterogeneous Hardware and Software Alliance.
NEW UPSCALE FRAMEWORK ALLOWS DEVELOPMENT OF REAL-TIME HPC APPLICATIONS
PSOCRATES, a project cofunded by the European Commission,
has created the UpScale Software Development Kit (UpScale
SDK) for the development of applications with strict timing and
highperformance requirements. The project members designed
and implemented an entirely new design framework, from
conceptual design of the system functionality to its physical
implementation, to facilitate the deployment of standardized
parallel applications in all kinds of realtime systems.
The UpScale SDK targets systems that demand more and more
computational performance to process large amounts of data
from multiple data sources, while requiring guarantees on
processing response times. Such systems may be used in
applications ranging from avionics to traffic management to the
stock exchange. Manycore processor architectures allow these
performance requirements to be achieved by integrating up to
hundreds of cores, interconnected with complex networks on
chip, paving the way for parallel computing.
Unfortunately, parallelization brings many challenges, by
drastically affecting the system’s timing behaviour. Providing
guarantees becomes harder, because the behaviour of the system
running on a multicore processor depends on interactions that
are not usually known by the system designer. This causes system
analysts to struggle to provide timing guarantees for such
platforms.
UpScale tackles this challenge by combining technologies from
different computing segments. These allow developers to
successfully exploit both the performance opportunities brought
by parallel programming models used in the highperformance
domain and timing analysis from the embedded realtime
domain, for the newest manycore embedded processors available.
HiPEACINFO 5124
Innovation Europe
The UpScale SDK includes the following components:
• Sourceto source compiler – enabling the analysis of source
code OpenMP parallelism annotations, extracting the required
information to allow for efficient and predictable mapping and
scheduling of parallel computations.
• Lightweight OpenMP tasking runtime – a smallfootprint, low
overhead implementation of the tasking model of the latest
OpenMP specification, which uses the information extracted by
the compiler to map OpenMP tasks to operating systems
threads.
• Embedded manycore operating system – a small kernel imple
mentation which efficiently handles parallel threads in many
core platforms, supporting both static assignment of threads to
cores as well as global scheduling approaches.
• Analysis tools – an integrated toolset for the timing and
schedulability analysis of realtime parallel applications.
The SDK framework is flexible and applicable to different use
cases, as demonstrated by the project results. Initially released
for the Kalray MPPA processor, it is possible to port to different
hardware architectures as also demonstrated by the project.
After a first showcase at the project’s Industrial Workshop which
took place in Porto in November 2016, the final results of the
project and the UpScale SDK were presented at Embedded World,
the international gathering for the embedded system technology
sector, which took place in Nuremberg in March 2017.
UpScale is openly available and released under commercially
friendly open source licences. It is available to download from
www.upscalesdk.com.
Contacts
Dr Luis Miguel Pinho, Project Coordinador ([email protected])
Dr Sandra Almeida, Project Manager ([email protected])
PROJECT: P-SOCRATES (Parallel Software Framework for Time-Critical
Many-core Systems)
START/END DATE: 01/10/2013-31/12/2016 (completed project)
KEY THEMES: high-performance, embedded, real-time, predictability
COORDINATOR: Instituto Superior de Engenharia do Porto (Portugal)
PARTNERS: Spain: Barcelona Supercomputing Centre, ATOS; Italy:
University of Modena and Reggio Emilia, Evidence SRL, Active
Technologies SRL; Switzerland: Swiss Federal Institute of Technology
Zurich (ETH Zurich)
BUDGET: €3.62M
WEBSITE: www.p-socrates.eu
P-SOCRATES received funding as part of the European Union's Seventh
Framework Programme (FP7/2007-2013) under grant agreement
no. 611016.
INNOHPC BRINGS HPC TO SMES IN THE DANUBE REGION
Launched on 1 January 2017, InnoHPC will enable enterprises as
well as academic and research institutions to cooperate more
closely in the field of supercomputing across national borders. In
the first stage, project partners will evaluate the current level of
exploitation of the regional highperformance computing (HPC)
infrastructures by small and medium enterprises. In the second
stage, they will design and establish a transnational InnoHPC
laboratory whose main purpose is to provide SMEs with remote
access to the HPC infrastructures. The third stage will involve a
pilot in which a number of SMEs from the electrotechnical and
automotive industry use the InnoHPC laboratory and web
platform.
The project runs until June 2019 and is supported by the Interreg
Danube Transnational Programme, whose priority is to promote
innovations and social responsibility in the Danube region. The
lead project partner is the Faculty of Information Studies in Novo
mesto (Slovenia). The focus is mainly on the eastern part of the
Danube region represented by the other partners in the project.
NAME: InnoHPC - High-Performance Computing for Effective
Innovation in the Danube Region
START/END DATE: 01/01/2017-30/06/2019
KEY THEMES: high-performance computing, industrial optimization
COORDINATOR: Faculty of Information Studies Novo Mesto (Slovenia)
PARTNERS: Czech Republic: IT4Innovations National Supercomputing
Center; Austria: RISC Software GmbH; Croatia: University of Rijeka;
Slovakia: Technical University of Košice; Romania: West University of
Timisoara, Executive Agency for Higher Education, Research,
Development and Innovation Funding; Slovenia: University of
Ljubljana, Electronic and Electrical Engineering Association – Chamber
of Commerce and Industry; Bulgaria: Research Centre for Regional
and Global Development; Hungary: Budapest Chamber of Commerce
and Industry
BUDGET: €2.05M
WEBSITE: http://bit.ly/Interreg_InnoHPC
Project co-funded by the European Union.
HiPEACINFO 51 25
Innovation Europe
INTRODUCING THE HETEROGENEOUS HARDWARE AND SOFTWARE ALLIANCE
Karim Djemame, University of Leeds and Oliver Barreto, Atos
In recent years we’ve seen the emergence of cyberphysical
systems (CPS) and the internet of things (IoT), promoted by
initiatives such as the European Commission’s Smart Anything
Everywhere, which have the potential to transform the way we
live and work. For example, the transformational impact of the
IoT in the long term is expected to increase significantly with
mass adoption and tens of billions of things being connected,
generating multitrillions of dollars in economic value. It will act
as a key driver behind new development platforms, analytics
applied to things and distributed/parallel architectures, as
Gartner highlighted in a 2015 report.
As the range of applications grows within CPS, the IoT, high
performance computing (HPC), mobile computing, wearable
computing, etc., there is an urgent need to design more flexible
software abstractions and improved system architectures to fully
exploit the benefits of the heterogeneous platforms on which
they operate. Heterogeneous parallel architectures have received
considerable attention for their ability to improve absolute
performance, minimize power consumption and/or lower costs
by combining different processor types in one system.
New platforms incorporating multicore CPUs, manycore GPUs
and a range of additional devices into a single solution are
increasingly being introduced. These platforms are emerging in a
wide range of environments, from supercomputers to personal
smartphones. One of the challenges for future application
performance lies not only with efficient nodelevel execution but
also with power consumption, a key focal point.
Complex engineering simulations often come to mind when
identifying families of applications that benefit most from
heterogeneous parallel architectures. However, in the upcoming
era of the IoT and big data there is significant interest in exploiting
the capabilities offered by customized heterogeneous hardware,
all of which have various memory hierarchies, size and access
performance properties.
Examples of customized heterogeneous hardware
FPGA – field-programmable gate array
ASIP –application-specific instruction set processor
MPSoC – multiprocessor system-on-chip
Heterogenous CPU (central processing unit) and GPU (graphics
processing unit) chips
Heterogeneous multi-processor clusters
In fact, online big data with nearly instantaneous results demand
massive parallelism and welldevised divideandconquer
approaches in order to exploit heterogeneous hardware, both
clientside and serverside, to its fullest extent. Heterogeneous
systems can handle workloads with fewer and/or smaller servers,
thereby saving costs. They can also slash the energy used to run
certain applications, which provides clear benefits in addressing
the growing interest in ‘green’ solutions and the pressure to
reduce the environmental impact of data centres, for example. As
the HiPEAC Vision notes, a common theme across all scenarios is
the need for lowpower computing systems that are fully
interconnected, selfaware, contextaware and selfoptimizing
within application boundaries.
Because the impact of hardware heterogeneity is rapidly
increasing, innovative architectures, algorithms and specialized
programming environments and tools are needed to efficiently
use these new and mixed/diversified architectures. For example,
the transitions to multicore processors, GPU computing and
hardwareasaservice (HaaS) cloud computing should be viewed
as a single trend. As the market for heterogeneous architectures/
multicore processors in embedded applications begins to move
into the product deployment stage, the need for software and the
underlying programming methodologies is also increasing in
parallel.
However, to put it bluntly, programming directly with this kind of
heterogeneity is a nightmare. We need to take action so that we
can rapidly develop solutions that help companies exploit
heterogeneous architectures to create richer experiences and
innovative business models. Those of us who are software
engineers need a mediator to help us deal with the heterogeneity
and take advantage of highlevel abstractions, as happened in
the past with traditional computing. We need to develop
innovative architectures, algorithms and even specialized
HiPEACINFO 5126
Innovation Europe
programming environments and tools to efficiently deal with
very different architectures in order to be more productive, obtain
rapid learning curves and to build solutions that are more general
purpose.
The TANGO project is creating a toolbox to provide the tools to
deal with this reality. However, we recognize that there are others
in the market and research arena who are also working towards
this direction. Our vision is to avoid creating yet more
heterogeneity in the tools and approaches that will lead this
market in the upcoming years. Therefore, instead of simply
pushing our own tools, TANGO aims to encourage collaboration.
We want to create something that will also create a link between
research and the market and exploit potential synergies.
To drive this forward, we have initiated the Heterogeneous
Hardware & Software Alliance (HHW&SW Alliance). The
initiative aims to unite the efforts of organizations interested in
the development of future technologies and tools to advance and
exploit computing and applications using heterogeneous
hardware. The alliance will focus on all phases of heterogeneous
hardware and software, from design time to enhanced execution,
parallel programming and optimized runtime. It will consider a
number of factors such as energy, performance, realtime, data
locality and security. This will enable a new way of developing
and executing nextgeneration applications.
The main aim of the alliance is to establish an organization in
which anyone interested in related technological areas can
collaborate. Participants will work towards a common objective:
founding a common, opensource and extendable set of
technologies and tools around the development for heterogeneous
hardware. These should be viable for mass adoption utilizing
technologies created by the alliance members, as well as being
attractive, easy to use and broader in scope and value. The
alliance is currently working on the establishment of a reference
architecture and is creating an online catalogue of tools and
technologies.
Interested in finding out more? Contact us: www.heterogeneityalliance.eu/contact
TANGO is funded by the European Commission under the H2020
Framework Programme for Research and Innovation under grant
agreement no 68758.
Further reading
TANGO: Transparent heterogeneous hardware Architecture
deployment for eNergy Gain in Operation
www.tango-project.eu
Smart Anything Everywhere
http://bit.ly/smart-anything-everywhere
Survey Analysis: The Internet of Things Is a Revolution Waiting
to Happen. Gartner report, January 2015
http://bit.ly/Gartner_Report15
HiPEAC Vision
www.hipeac.net/publications/vision
HiPEACINFO 51 27
Peac performance
Rosa M Badia, Barcelona Supercomputing Center (BSC)
Superscalar architectures became
popular in the early 1990s and are
still in use in most processors. A
superscalar processor implements
instructionlevel parallelism and can
execute several instructions in a
clock cycle by trying to feed all the
execution units on the processor.
Traditionally, computer architecture instructions are programmed
sequentially, and a program is an ordered list of instructions.
However, designs such as the superscalar processor allow parallel
execution, with several instructions being executed at the same
time. What is more, the data dependencies between instructions
are dynamically detected by the CPU. This means that outof
order execution and even speculative execution – where some
instructions are executed that may not actually be needed – are
possible. All these techniques help make the most of the
instruction cycle and hence minimize delays in processing.
In the superscalar programming model we follow the same
strategy, but implemented at software level. PyCOMPSs is a task
based programming model that offers an interface based on
Python sequential code, but that enables execution in parallel by
building a datadependency graph of the application tasks at
execution time. A task in PyCOMPSs is a method or a function,
annotated with a Python decorator which also describes the
directionality of the task parameters.
The directionality of a parameter can be ‘in’ when a parameter is
read, ‘out’ when a parameter is written, and ‘inout’ when a parameter
is both read and written by the task.
Tasks in PyCOMPSs are equivalent to the instructions in super
scalar processors, and the objects or files exchanged by the tasks
are equivalent to the registers in the processor. The directionality
of the parameters is used to derive data dependencies between
instances of tasks at execution time. Another idea borrowed from
the computer architecture world is register renaming, which we
apply to the objects/files exchanged by the tasks.
The syntax of PyCOMPSs is minimal, using decorators to annotate
tasks and a small API for synchronization. PyCOMPSs relies on a
runtime that is able to exploit the inherent parallelism at task
level and to execute the application on a distributed parallel
platform (clusters and clouds) while the code remains agnostic
of the existing hardware. The runtime is responsible for scheduling
the tasks on the available computation resources, performing the
necessary data transfers between distributed memory spaces,
synchronizing all activities and interoperating with heterogeneous
computing resources including cloud middlewares, etc.
Parallelizing Python codes using the superscalar paradigm
Rosa Badia of Barcelona Supercomputing Center explains how PyCOMPSs takes inspiration from computer architecture to enable
execution in parallel at the software level.
HiPEACINFO 5128
Peac performance
Parallelizing Python codes using the superscalar paradigmPython was chosen rather than other programming languages
due to its popularity in several scientific communities. The high
level of adoption is not only due to the language features, but
also because of the large number of thirdparty libraries available
for the community. Examples of very popular libraries are NumPy
or SciPy, which offer data structures and numerical routines.
NumPy automatically maps operations on vectors and matrices
to the functions of BLAS and LAPACK numerical libraries, which
are highly optimized and in many cases parallelized at
threadlevel.
With this approach, in PyCOMPSs we can exploit two levels of
parallelism: tasklevel, between nodes of a distributed computing
platform, and threadlevel, inside the nodes. Results with
numerical kernels such as matrix multiplication, Cholesky or QR
factorizations show that we can get performances of up to 60% of
the peak in one node and very good scalabilities up to thousands
of cores. In all cases, it is important to take into account that the
user application is a sequential annotated Python code that is
executed in parallel by the PyCOMPSs runtime.
With the objective of giving support to new storage technologies
such as NVRAMs or SSDs, PyCOMPSs has been enabled to
support persistent storage. The programmer is able to define
which objects of their program will remain persistent after the
execution of the code, although access to these objects by the
program remains unchanged. Similarly, programs can access
objects that were made persistent before. This enables very easy
support of producerconsumer sets of applications, or insitu
processing, visualization, etc. This support has been made
possible by the definition of a Storage API with a set of methods,
some accessed by the PyCOMPSs runtime and others offered
directly to the programmer. The interface for the programmer is
minimal, with a method to make the objects persistent and with
constructors to access data that is already persistent.
Example of how objects can be persisted with the storage API. The object X is created and made persistent. Later, it is accessed normally in the loop
Two backend solutions to store the persistent storage are currently
supported: Hecuba and dataClay. Hecuba is a set of tools and
interfaces that aims to provide programmers with an efficient
and easy interaction with nonrelational databases. More
specifically, Hecuba implements an interface to access data stored
in Cassandra databases, such as regular Python memory objects.
The objectoriented approach is focused by dataClay, a data store
that offers novel techniques for sharing structured data in a
multiprovider ecosystem.
While the PyCOMPSs runtime operates well with clusters and
clouds, up until now it has not been enabled for GPUs, FPGAs or
other types of accelerator. To support these types of architecture,
a hybrid programming model that combines PyCOMPSs, to deal
with coarsegrain tasks, and the BSC programming model
OmpSs, to deal with finer granularity, is under research. This and
other challenges of taskbased programming models will be
described and discussed in my lectures at the ACACES summer
school this year.
COMPSs is open source. Code, compiled packages and documentation are available at compss.bsc.es
Rosa Badia is delivering the course ‘Application programming on parallel/distributed computing platforms’ at ACACES17. Further information: http://bit.ly/ACACES17_Rosa_Badia
“Tasks in PyCOMPSs are equivalent to the instructions in superscalar processors, and the objects or files are equivalent to the registers”
HiPEACINFO 51 29
Tech opinion
Neuromorphic computing explores largely
uncharted territory in several sciences of
living and artificial systems. The idea of
different kind of computing was driven by
the extraordinary capabilities of human
brain to perform on very low levels of
energy consumption. As physical limita
tions began to reach barriers which cannot
be exceeded without drastically changing
the underlying architecture, some 30 years
ago scientists and engineers started thinking
about how to emulate human brain
performances with neuro morphic chips.
Looking at the big picture of computation
vs. communication types of signals (see
figure 1), one should think about neuro
morphic computing as digital communi
cation and analogue compu tation – very
much as the human brain works. Analogue
computing preceded digital computing,
which is dominant today and reaching its
limits. The objective is to achieve the
energy efficiency of the human brain,
which consumes only 20 watts as com
pared with contemporary HPC systems
which consume megawatts of electricity.
Figure 1: Computation vs communication signals
Traditional digital computer technology is
slated to deliver an exascale machine
(1018 FLOPs with 1018 bytes of memory),
but delivering an order of magnitude more
performance will require a radical change
of computing paradigm. This change will
have an impact on the entire chip
hardwaresoftware stack (figure 2) and
imply not only novel neuromorphic
devices but also different types of software
running new types of algorithm. In short,
an entirely new technology landscape and
architecture.
Figure 2: Exascale calls for the reinvention of the full stack
An interesting consequence of this will be
that the different types of problem can be
addressed in a different way and exhibit
superior performances (figure 3). For
example, the use of quantum computing
and highperformance computing to
design effective neuromorphic computer
architectures is being investigated. It is
not beyond the realms of imagination that
an entire new age of computing might
open up.
Figure 3: Different problems may require different approaches
In conclusion, as we witness huge invest
ment in human brain research in the Euro
pean Union and the United States,
amoun ting to the investment of one billion
dollars/euros over the next decade, one
should not make the mistake of thinking
that the human brain will be replaced with
artificial systems. The big challenges will be:
1. To construct systems which consume
watts instead of megawatts of electricity.
2. To reinvent fault tolerance so that it
resembles that of the human brain,
which is constantly improving while
losing neurons daily.
3. To construct systems that can learn from
massive interactions instead of being
programmed.
New computing paradigm(s) will open up
splendid opportunities for the next gene
ration of scientists and engineers con
structing entirely new systems and
deploying novel devices across the entire
architecture stack.
Neuromorphic Computing: low-power systems, the brainy way
Attempts to recreate the human brain in silico have sparked fears that artificial systems will replace the human brain. This is a misconception,
argue Kemal Delic and Dave Penkler of Hewlett Packard Enterprise. Instead, the human brain – the original computer – is the inspiration
for next-generation computing systems.
Figures in this article are adapted from the following sources:Rebooting Computing: The Road Ahead by T.M. Conte et alhttps://www.computer.org/csdl/mags/cs/2017/02/mcs2017020014.htmlA Neuromorph’s Prospectus by K. Boahenhttps://www.computer.org/csdl/mags/co/2017/01/mco2017010020.html
HiPEACINFO 5130
Tech Entrepreneur
Technological innovation has a history of
changing the world in major ways and
making a real impact on society. Just think
of the British Agricultural Revolution of
1700 and the Industrial Revolution of
1780 – they were all about changing the
basic way in which people worked and
produced things by innovatively applying
and scaling the technology breakthroughs
of the time. We have been in a compu
tationdriven revolution for some decades;
it drives much of the modern economy
and compe titiveness for the countries in
which we live. How do we keep the
revolution going?
For those currently doing interesting
scientific research, it is a good idea to take
some time to look at how their findings
might be used by companies and indivi
dual consumers: seeing someone using
the results of your labours and insights is
just as satisfying as getting a best paper
award. It can be a lot of fun, evange lizing
the technology area you enjoy and having
a wider group of people understand and
appreciate it more. It can also be financially
rewarding.
A vision in silicon
Tech entrepreneur Peter Denyer invented
the silicon optical sensor, now used in
most digital cameras. He founded VLSI
Vision and sold it to ST Microelectronics,
before founding a number of other
companies.
Europe has a very good record of making
scientific breakthroughs and though we
have made great strides in turning these
into products people and companies find
useful, it is an area in which we ought,
and need, to do better. We have tended to
lag behind the USA in this dimension. A
number of reasons have been put forward
as to why this has been the case:
We do not have as large a single market
as the USA
That ought not to be true if we take the
European Union (EU) single market into
account. In the digital domain a global
market is accessible instantly through the
internet. As we look at the geographic
addressable market we do need to
consider the wider global market – not
just our individual country.
The USA has easier access to venture
capital and investment
This is no longer the case with a number
venture funds and angel investment
groups now active throughout Europe. As
of the end of 2016 we have 47 unicorns
(startup companies with a valuation
>$ 1 billion) in Europe, with an average
investment of some $ 260 million each. So
there is investment capital available to be
won.
The USA has more technical talent and
knowhow
Europe has a huge talent pool both in the
technical and entre preneurial domain.
Keep the revolution going
Thinking about converting your research results into market-ready products and services? As Director of Commercialisation at the University of Edinburgh from 2006 to 2016, Colin Adams presided over the creation of 70 spin-outs and start-ups which collectively raised over e 250 million in equity investment. Here he explains why and how you should be a technology entrepreneur.
HiPEACINFO 51 31
The strength of our research and univers
ities is key. There is also a growing band of
seasoned technical entrepreneurs who
already have at least one success under
their belt and are available to mentor and
help new comers.
So there is no reason why we should not
make even more headway in turning the
worldleading research that this
community has been carrying out into
innovative products.
Improbable sums
The founders of the UK virtual reality
start-up Improbable Worlds, Herman
Narula, Rob Whitehead and Peter Lipka,
recently raised $520 million from Softbank.
The company is now reported to be valued
at over $1 billion. Between them, the three
founders hold more than 50% of the equity.
If you do have some technology you think
you would like to commercialize, first
work up a very simple working prototype
of what you think people would use. Find
a couple of potential customers (preferably
not coresearchers) and try it on them so
you get some feedback. Listen to their
input and modify your approach based on
that. The prototype can also be used to
show potential investors what you have in
mind, and to people you may want to
recruit onto your team.
The other area to think through is how
you protect the intellectual property in
your invention, e.g. by patenting some
element of it, so others don’t capitalize on
your efforts without you gaining anything
from it.
If you are looking at seriously pursuing
this route, I would advise seeking out
some business mentors who have solid
business experience and contacts, so you
have people to provide feedback on your
ideas for a business plan and how to
develop the commercial side.
An interesting thought in the space we
address in the HIPEAC community is what
happens next to keep Moore’s law – and
with that the technology revolution –
moving forward. I had an interesting
conversation with Steve Jurvetson recently.
He is one of Silicon Valley’s most successful
investors and he came up with the 120
year view of Moore’s Law shown below.
Now, does your research give guidance to
how that curve continues? If so, you could
have something very valuable – so start
thinking about how to commercialize it.
Colin Adams is delivering the Technology Innovation and Entrepreneurship course at the HiPEAC summer school ACACES17.http://bit.ly/ACACES17_ColinAdams
Tech Entrepreneur
HiPEACINFO 5132
As computing systems evolve towards
complex multicore designs with deep and
diverse memory hierarchies, improving
the performance and optimizing the exe
cution of realworld applications become
of fundamental importance. In high
performance computing environ ments, it
is crucial, but not trivial, to determine
which hardware resources represent the
main execution bottlenecks that limit the
application performance, especially when
deciding on the most adequate software
optimization technique to be applied.
To support this decision process, Aleksandar
Ilic, Frederico Pratas and Leonel Sousa,
researchers from INESCID, Instituto
Superior Técnico, University of Lisbon, and
members of HiPEAC, proposed a set of
fundamental Cacheaware Roofl ine
models, which provide a simple and
intuitive way of visually representing the
limits of parallel processing on contem
porary multicore processors.
These models evaluate how key micro
archi tectural aspects, such as accessing
diffe rent functional units or different
memory hierarchy levels, affect realistically
achie vable upperbounds for performance,
power consumption and energyefficiency
on a given multicore architecture. They
have been used to characterize the
behaviour and improve the efficiency of
several realworld applications, e.g. in the
areas of scientific computations and
bioinfor matics.
In 2017, a team of Intel software deve
lopers (leaded by Zakhar Matveev, Roman
Belenov and Philippe Thierry) successfully
integrated the performance Cacheaware
Roofline model as an official feature of
Intel® Advisor, which is part of the Parallel
Studio XE suite (Intel’s main application
development frame work). Within Intel®
Advisor, the process of building the roof
line plots and indepth application charac
te rization are fully automatized with
res pect to the hardware platform where
the applications are executed. The support
for a wide range of Intel devices is also
provided, which covers all contempo rary
Intel CPU microarchi tectures (from
Nehalem to Skylake) up to massively
parallel devices (e.g., Intel Xeon Phi
Knights Landing).
A brief overview of the Cache-aware Roofline in Intel® AdvisorThe performance Cacheaware Roofline is
plotted with the X axis as Arithmetic
Industry focus
Cache-aware Roofline Model in Intel® AdvisorA bird’s eye view for parallel processing
Leonel Sousa, Aleksandar Ilic’ and Frederico Pratas of the University of Lisbon describe how they developed a simple way to visualize the limits
of parallel processing, which has been integrated into Intel® Advisor.
“The models evaluate how key micro-
architectural aspects affect upper-bounds for
performance”
HiPEACINFO 51 33
Intensity (measured in FLOPs/Byte) and
the Y axis as the performance in GFLOPs/
Second, both in logarithmic scale. Before
collecting data of a specific application,
Intel® Advisor automatically runs a set of
quick benchmarks to measure the hard
ware limitations of the used processor,
which it then plots as lines on the chart,
called roofs (see Figure 1). The horizontal
lines represent the number of floating
point computations (of a given type) that
the underlying hardware can perform in a
given span of time. The diagonal lines are
representative of how many bytes of data
a given memory hierarchy level can
deliver per second.
Each dot represents a loop or function in
the program, and its position in the
Roofline plot indicates performance and
Arithmetic Intensity. The size and colour
of the dots in Intel® Advisor’s Roofline
chart indicate how much of the total
program time a loop or function takes.
Small, green dots take up relatively little
time, so are likely not worth optimizing;
large, red dots take up the most time, so
they are the best candidates for optimi
zation, especially the ones with a large
gap to the topmost attainable roofs. In
general, the farther a dot is from the
topmost roofs, the more room for improve
ment there is. For example, the Scalar Add
Peak represents the maximum possible
performance without taking advantage of
vectorization, as indicated by the next
roof up being the Vector Add Peak.
Where can I get Intel® Advisor with Cache-aware Roofline?As stated in the Intel early access pro
gramme, ‘the Intel Advisor offers a great
step forward in memory performance opti
mi zation with a new vivid Advisor
“Roofline” bounds and bottlenecks ana
lysis’. Cacheaware Roofline is currently a
feature of Intel® Advisor beginning
officially with version 2017 Update 2,
which is part of the Parallel Studio XE suite
(Cluster Edition and Professional Edition).
Cache-aware Roofline in Intel® Advisor
‘The Intel Advisor offers a great step forward in
memory performance optimization with
a new vivid Advisor “Roofline” bounds and
bottlenecks analysis’
Further reading
A. Ilic, F. Pratas, and L. Sousa. ‘Cache-aware Roofline model: Upgrading the loft.’
IEEE Computer Architecture Letters, vol. 13, n. 1, pp. 21-24, 2014.
‘Beyond the Roofline: Cache-aware Power and Energy-Efficiency Modeling for Multi-cores.’
IEEE Transactions on Computers, vol. 66, n. 1, pp. 52-58, 2017.
Intel. (2017) Intel® Advisor Roofline
http://bit.ly/Intel_Advisor_Roofline
Intel. (2017) Intel Parallel Studio XE
http://bit.ly/Intel_Parallel_Studio_XE
Industry focus
HiPEACINFO 5134
COMPANY: Rimac Automobili
MAIN BUSINESS: Creating next generation
electric hypercars; providing full solutions
to global original equipment manufacturers,
from high-performance electric vehicle
components to turnkey solutions
LOCATION: Sveta Nedelja, Croatia
WEBSITE: www.rimac-automobili.com
The seeds for the company were first planted back in 2007, when
Mate Rimac started tinkering with his BMW E30 as a hobby, after
the car’s engine blew up during a race. When he managed to give
it an electric powertrain – the engine and transmission – the
project soon attracted press interest. In total, Mate broke five
Fédération Internationale de l’Automobile and Guinness World
Records for the fastest accelerating electric car with this
prototype, affectionately known as the ‘Green Monster’. Wanting
to develop a more powerful version, Rimac brought together a
team of experts to develop their own components, in order to
deliver genuinely highperformance electric propulsion. Thanks
to selling patents and gaining financing from angel investments,
Rimac Automobili was founded in 2009.
The ‘Green Monster’ was the original Rimac prototype
Lovingly crafting each car in house – including everything from
the battery to the transmission to the infotainment system – the
team built on their knowledge from developing the BMW E30
prototype to create first the Concept_One and then the Concept_S.
Accelerating from 0100km/h in just 2.5 seconds, the cars have
beaten even top supercars in test races.
Check out the Concept_One beating LaFerrari in this video:
http://bit.ly/Rimac_ConceptOneLaFerrari.
The Rimac Concept One has beaten LaFerrari
Rimac is fortunate to benefit from a great team and we’ve been
working with the education system in Croatia to help prepare
engineers with the requisite skills. As a smaller company, at
Rimac you have the satisfaction of working on the whole lifecycle
of the product, from design, to production, to testing. This is
contrast to working at a large multinational where you would
likely be working on one small part of the development. Rimac is
currently looking for topquality staff, with roles encompassing
everything from battery engineering to embedded engineering.
We are seeking software engineers for embedded systems
development, from the initial idea to the final product. Candidates
should have C/C++ programming skills, familiarity with basic
electronic circuits and basic knowledge of communication
protocols. Experience with realtime operating systems and
embedded systems projects experience are more than a plus. We
are also recruiting an embedded hardware engineer, who should
have knowledge of electronic components and systems, the
programming language C and communication protocols (UART,
SPI, I2C, CAN...).
Interested in working at Rimac Automobili? Check out the vacancies on
the HiPEAC Jobs Portal: hipeac.net/jobs
SME snapshot
The drive to succeed
From hobby project to building the world’s fastest electric car, Rimac Automobili is a high-octane success story. Here, Chief Human Resources
Officer Aco Momcilovic’, Director of Components Matija Gracin and Marketing Assistant Marta Longin show how this company punches
above its weight and give an insight into why it is a great place to work.
HiPEACINFO 51 35
HiPEAC futures
What first got you interested in computer
science?
My interest began a long time ago. I think my
career started with the first computer my father
bought me. It was a very small Texas Instru
ments (TI) processorbased system which was
programmed with hexadecimal code. Later I
had a TI 99/4A and then of course a Commodore
C64. I was born at a time when computer
systems developed very quickly in a very short
period of time, a phenomenon which has
continued up to the present day. A fascination
with novel systems, processor system trends
and energy optimization has kept my interest
all this time.
What makes embedded computing an exciting
field to work in?
I think embedded systems is the most important
domain, both for industry and academia. We
are seeing trends where embedded technology
is even entering the realm of highperformance
computing due to the effectiveness of perfor
mance per watt. Trends like cyberphysical
systems (CPS) and the internet of things (IoT)
enable better processor systems in terms of
ultralow power and normally high realtime
capability. This is leading to changes in many
fields. Embedded systems form the ‘motor’
which drives megatrends in machine learning,
CPS and IoT.
The TI 99/4A was one of Michael’s first computers Photo credit: Manolis Klaglas, Flickr
creativecommons.org/licenses/by-nc-sa/2.0
Can you describe some highlights of your
career so far?
One highlight in my career was my decision to
study. I was doing a job in the electronics field
and came to a point where I needed to choose
to study or stay on the path I was on. It was not
easy for me because I gave up a permanent
position in order to embark on the ‘adventure’
of studying. What made it even more daunting
was that, at the time, there were no guarantees
that it would become a job. Later, after finishing
my diploma, it was thanks to my former PhD
advisor believing in me that all other steps
were possible. Now, I have the best job I could
wish for as a professor at RuhrUniversität
Bochum .
What makes Europe a great place to work?
I think Europe is an excellent place to work, as
there is a network of excellent researchers and
exchange of knowledge and experience is easily
achieved. In my group at Bochum I always
have researchers from Europe, but also from all
over the world. We like to discuss issues with
other researchers and Europe is an excellent
platform to do so. Funding oppor tunities,
mostly the European Commission ICT calls, are
often an especially good way to collaborate.
What advice would you have for researchers
embarking on their careers now?
I think the most important thing is to stay
curious and open minded, and never be
restrictive with sharing knowledge. Teamwork
is important, as is supporting activities like
EUROPRACTICE (www.europracticeic.com)
which offers microelectronic and micro system
design, prototyping, production and test
services. EUROPRACTICE is an extremely
important source of support for researchers in
Europe and I hope it continues for a long time.
Michael Hübner is delivering a course titled ‘Reconfigurable Hardware, Tools and Applica-tions’ at ACACES 2017http://bit.ly/ACACES17_MichaelHuebner
Career talk: Michael Hübner, Ruhr-Universität Bochum
HiPEACINFO 5136
HiPEAC futures
Creating the future through international exchange: HiPEAC collaboration grants
Collaboration grants allow PhD students and junior post-doctoral researchers within the HiPEAC network to work jointly with a new research group to work
on key challenges for computing systems. For further information about how to apply, visit www.hipeac.net/mobility/collaborations.
NAME: Joshua Lant
RESEARCH CENTRE: University of Manchester
HOST INSTITUTION: Foundation for Research
and Technology Hellas
DATES OF COLLABORATION:
26/07/2016 - 28/10/2016
Epic memories: When Manchester met ‘Silicon Island’, Crete
The Horizon2020 ExaNeSt project is developing and prototyping
network, storage and cooling solutions to make exascale com
puters – those capable of a billion billion calculations per second
– feasible. To help achieve this, ExaNeSt uses UNIMEM: a form of
global address space that allows multiple boards to share memory
regions between them and to communicate at low overhead and
low energy consumption. UNIMEM was developed during the
EUfunded EUROSERVER project, which ended earlier this year.
I spent three months at FORTH to learn about the technical
details of the UNIMEM memory model, its limitations and its
implementation on the Xilinx Ultrascale+ architecture. During
this time I also investigated the requirements of the interconnect
to properly allow the use of UNIMEM. In addition, I worked to
advance the design on FPGA of a custom AXIbased interconnect
for use in the ExaNeSt project at the daughtercard level, and
implementation and testing of this design on Ultrascale+
hardware at FORTH.
This collaboration resulted in documentation on the details of
UNIMEM, with particular focus on the ways in which the memory
model affects the programmability of the hardware and the
differences between the ideal UNIMEM model and the current
implementation. Using the Vivado toolchain, we also created
hardware designs which measure the performance of the custom
AXI interconnect with traffic generators. This has been
implemented and run successfully on the Ultrascale+ hardware.
Real performance metrics can now be gathered and the design
improved to meet higher performance constraints, with more
interesting features of the interconnect design being added and
tested. The aim is to publish the results of the implementation
and performance characteristics of these more advanced features.
The scope for future collaboration is very wide, considering that
ExaNeSt works closely together with the ExaNoDe and EcoScale
projects, which will all run until late 2018, and then the new
EuroEXA project will continue along the same lines until early
2021. Developing networks with the people at FORTH has been
highly beneficial. Now that I’ve spent time with specific members
of the team with specific skills at FORTH, it will be much simpler
to find avenues where FORTH and Manchester may assist one
another on specific parts of the project as work progresses.
Manolis Katevenis of FORTH commented: ‘Joshua’s visit was an
important step in furthering our collaboration with the University
of Manchester, especially on Interprocessor Communication
protocols – a key component for the Modularity of Silicon that
future technologies demand.’
He added: ‘This collaboration builds on FORTH’s strong tradition
of working with colleagues across Europe in R&D projects, which
has already yielded great results. Last year, for example, FORTH
attracted to Crete a Development Lab of the startup KALEAO,
cofounded by University of Manchester professor John Goodacre.
We look forward to continue this strong relationship to help
deliver the exascale machines that Europe desperately needs.’
Dubbed ‘Silicon Island’, Crete offers great opportunities for young computer scientists
HiPEACINFO 51 37
HiPEAC futures
HiPEAC internships: your career starts here
Every year, HiPEAC company members propose topics for internships, which are then funded or part funded by the project. This year, there
was a record number of internships and applicants. For further information, visit the HiPEAC website: hipeac.net/mobility/internships.
In his internship, Hamzeh Ahangari helped Embedded Computing Specialists lay the groundwork for a safety-qualified embedded system
in the transport domain.
NAME: Hamzeh Ahangari
RESEARCH CENTRE: Bilkent
University
HOST COMPANY: Embedded
Computing Specialists
DATES OF INTERNSHIP:
06/11/2016 - 03/02/2017
Transport safety as standard
Safetycritical computers are used exten
sively in many civil domains, such as the
automotive sector, rail transport and
avionics. While general, safe micro con
trollers with limited processing capabi
lities are available on the market, safe
processors with intensive processing capa
bilities should be designed and tailored to
the specific application requirements.
Such designs are always based on com
ponentlevel or systemlevel redundancies.
During this internship, we did a prelimi
nary study of the required standards for
implementing safetycritical electronic
systems. There are several such standards,
including IEC61508, DO178, etc, each of
which addresses a particular domain. From
these documents, the main para meters
addressed were identified and investigated.
These include component failure rates,
common cause failure rate, diagnostic
coverage factor, repair rate and so on.
The second step was to focus on choosing
the platform. Unlike existing systems,
which are usually implemented in
commer cialofftheshelf processors, like
those from Texas Instruments or NXP, we
wanted to check the feasibility of imple
menting such systems on highperfor
wwmance FPGAs like Xilinx Zynq. We
undertook a thorough investigation of the
facilities that FPGA manufacturers provide
for reaching the standards’ safety levels.
We also implemented a draft version of a
compact printed circuit board for Xilinx
Zynq FPGA.
The main aim was to assess the feasibility
of designing and manufacturing a safety
qualified embedded system for the
transportation industry. With the prelimi
nary study phase and part of the FPGA
system implementation complete, I hope
that we will be able to continue working
together on such systems in future.
Embedded Computing Specialists cofoun
der Philippe Manet commented: ‘This
HIPEAC internship allowed us to quickly
understand key aspects of certification in
our business, thanks to the expertise and
previous experience of Hamzeh. It allowed
us find the right strategy to handle this
complex issue.’
“The main aim was to assess the feasibility of
designing and manufacturing a safety-qualified
embedded system for the transportation
industry”
HiPEACINFO 5138
HiPEAC futures
Three-minute thesis
Continuing our series highlighting the work of HiPEAC’s 800-strong network of PhD students, in this issue we focus on memory controller
architecture for real-time applications.
NAME: Yonghui Li
RESEARCH CENTRE: Eindhoven University of
Technology
ADVISERS: Professor Kees Goossens and
Dr Benny Akesson
THESIS TITLE: Design and Formal Analysis of
Real-time Memory Controllers
Featured research: Memory access control which is right on time
Heterogeneous multicore computing platforms are becoming
increasingly popular to achieve the necessary computational
power without using too much energy. These platforms
simultaneously execute multiple applications that interact with
the physical world. The applications typically have realtime
requirements, such as a time deadline or a throughput constraint,
which must be satisfied to ensure the safety or performance of
the applications.
However, it is very challenging to satisfy the realtime
requirements of each application because of the complex
interference caused by resource sharing in the platform. In
particular, synchronous dynamic randomaccess memory
(SDRAM) stores the data and instructions in the platform and is
accessed by other resources issuing read or write memory
requests. It is one of the most shared resources and hence has
great impact on the timing behavior of applications.
This dissertation addresses two critical issues: 1) how to
efficiently serve memory requests with variable sizes, which are
generated by heterogeneous resources in the platform, and 2)
how to guarantee the worstcase response time and bandwidth
when accessing SDRAM, such that application requirements can
be satisfied. A new memory controller architecture is proposed to
serve the variablesized requests by dynamically scheduling the
proper number of memory commands to SDRAM at runtime
(RunDMC). Thanks to the dynamic scheduling scheme, the
memory commands for successive requests are pipelined,
resulting in higher efficiency than static command schedules for
requests. To guarantee the worstcase response time and
bandwidth, three analysis approaches are presented, including a
mathematical formalization, a dataflow model, and a timed
automata model, respectively. These models have different
properties that support the simulation, validation and verification
of the memory controller.
This dissertation is the first to explore and compare different
analysis approaches to obtain the worstcase response time and
bandwidth of realtime memory controllers. The experimental
results demonstrate that the proposed dynamicallyscheduled
memory controller is more efficient than existing (semi)static
controllers for requests with variable sizes, and the derived
worstcase results are tightly bounded.
Models Worst-Case Analyses Tools
Timed Automata model (TA)
Mode-controlled dataflow model
(MCDF)
Mathematical model
Analytical approach (ALAP & Collision)
Scheduled approach (ALAP)
MCDF Analysis
(Collision)
TA analysis
RTMemCtrl: Open source RTMemCtrl: Open source
Ericsson Heracles Ericsson Heracles
SystemC Simulator SystemC Simulator
Cycle-accurate model
Validation Accuracy equal
low high Worst-case
witness Correspondence
HiPEACINFO 51 39
25 July 2017
HPCA 2018 (Vienna) abstract submissions
HiPEAC Paper Award conference
hpca2018.ece.ucsb.edu
2017 JULY
2017 AUGUST
2017 SEPTEMBER
Dates for your diary
4 August 2017
ASPLOS 2018 (Williamsburg, VA, USA)
abstract submissions
HiPEAC Paper Award conference
asplos2018.org/calls
11 - 13 September 2017
2017 ARM Research Summit
Cambridge, UK
developer.arm.com/research/summit
15 September 2017
PDP 2018 (Cambridge, UK)
paper submissions
bioinfo.itb.cnr.it/pdp2018/call.html
4 - 8 September 2017
FPL 2017
Ghent, Belgium
fpl2017.org
28 August - 1 September 2017
Euro-Par 2017
Santiago de Compostela, Spain
europar2017.usc.es
Phot
o: J
L C
ernad
as, F
lick
rcr
eati
veco
mm
ons.
org/l
icen
ses/
by/2
.0
12 - 15 September 2017
ParCo 2017
Bologna, Italy
parco.org
Phot
o: Y
uri V
irov
ets
crea
tive
com
mon
s.or
g/l
icen
ses/
by/2
.0
6 - 8 September 2017
ACM Europe Conference 2017 ft. HiPEAC
Barcelona, Spain
acmeurope-conference.acm.org
Phot
o: Z
apri
ttsk
y, F
lick
rcr
eati
veco
mm
ons.
org/l
icen
ses/
by/2
.0