Presentación a cargo de Mateo Valero, Director del Barcelona Supercomputing Center, en el marco de la 30ª edición de los Encuentros de Telecomunicaciones y Economía Digital.

1. ¿Es posible construir
el Airbus de la
Supercomputación
en Europa?
Prof. Mateo Valero
BSC Director
05/09/2016

2. Technological Achievements
• Transistor (Bell Labs, 1947)
• DEC PDP-1 (1957)
• IBM 7090 (1960)
• Integrated circuit (1958)
• IBM System 360 (1965)
• DEC PDP-8 (1965)
• Microprocessor (1971)
• Intel 4004

3. The MultiCore Era
Moore’s Law + Memory Wall + Power Wall
Chip MultiProcessors (CMPs)
UltraSPARC T2 (2007)
Intel Xeon
7100 (2006)
POWER4 (2001)

4. Top 500 Supercomputers - June 2016
Rank Name Site Computer Total
Cores Rmax Rpeak Power Mflops/
Watt
1
Sunway
TaihuLight
National Supercomputing Center
in Wuxi
Sunway MPP, Sunway SW26010
260C 1.45GHz, Sunway
10649600 93014593,88 125435904 15371 6,305
2
Tianhe-2
(MilkyWay-2)
National Super Computer Center
in Guangzhou
TH-IVB-FEP Cluster, Intel Xeon
E5-2692 12C 2.200GHz, TH
Express-2, Intel Xeon Phi 31S1P
3120000 /
2736000
33862700 54902400 17808 1901,54
3 Titan
DOE/SC/Oak Ridge National
Laboratory
Cray XK7 , Opteron 6274 16C
2.200GHz, Cray Gemini
interconnect, NVIDIA K20x
560640 /
261632
17590000 27112550 8209 2142,77
4 Sequoia DOE/NNSA/LLNL
BlueGene/Q, Power BQC 16C
1.60 GHz, Custom
1572864 17173224 20132659,2 7890 2176,58
5
RIKEN Advanced Institute for
Computational Science (AICS)
K computer, SPARC64 VIIIfx
2.0GHz, Tofu interconnect
705024 10510000 11280384 12659,89 830,18
6 Mira
DOE/SC/Argonne National
Laboratory
BlueGene/Q, Power BQC 16C
1.60GHz, Custom
786432 8586612 10066330 3945 2176,58
7 Trinity DOE/NNSA/LANL/SNL
Cray XC40, Xeon E5-2698v3 16C
2.3GHz, Aries interconnect
301056 8100900 11078861
8 Piz Daint
Swiss National Supercomputing
Centre (CSCS)
Cray XC30, Xeon E5-2670 8C
2.600GHz, Aries interconnect ,
NVIDIA K20x
115984
73808
6271000 7788852,8 2325 2697,2
9 Hazel Hen
HLRS - Höchstleistungsrechenzentrum
Stuttgart
Cray XC40, Xeon E5-2680v3 12C
2.5GHz, Aries interconnect
185088 5640170 7403520
10 Shaheen II
King Abdullah University of
Science and Technology
Cray XC40, Xeon E5-2698v3 16C
2.3GHz, Aries interconnect
196608 5536990 7235174 2834 1953,77

5. Performance Development of HPC
over the Last 23 Years from the Top500
0,1
1
10
100
1000
10000
100000
1000000
10000000
100000000
1E+09
1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
59.7 GFlop/s
400 MFlop/s
1.17 TFlop/s
93 PFlop/s
286 TFlop/s
567 PFlop/s
SUM
N=1
N=500
1 Gflop/s
1 Tflop/s
100 Mflop/s
100 Gflop/s
100 Tflop/s
10 Gflop/s
10 Tflop/s
1 Pflop/s
100 Pflop/s
10 Pflop/s
1 Eflop/s

6. J. Dongarra
Performance Development of HPC over the Last
23 Years from the Top500
59.7 GFlop/s
400 MFlop/s
1.17 TFlop/s
33.9 PFlop/s
166 TFlop/s
362 PFlop/s
SUM
N=1
N=500
1 Gflop/s
1 Tflop/s
100 Mflop/s
100 Gflop/s
100 Tflop/s
10 Gflop/s
10 Tflop/s
1 Pflop/s
100 Pflop/s
10 Pflop/s
1 Eflop/s
My Laptop 70
Gflop/s
My iPhone 4 Gflop/s

7. The Evolution of the Research Paradigm
Numerical
Simulation and
Big Data Analysis
• Reduce expense
• Avoid suffering
• Help to build knowledge where
experiments are impossible or
not affordable

8. HPC: An enabler for all scientific fields
Advances leading to:
• Improved Healthcare
• Better Climate Forecasting
• Superior Materials
• More Competitive Industry
Life Sciences
& Medicine
Earth
Sciences
Astro,
High Energy
& Plasma
Physics
Materials,
Chemistry &
Nanoscience
Engineering

9. Barcelona Supercomputing Center
Centro Nacional de Supercomputación
BSC-CNS objectives:
• Supercomputing services to
Spanish and EU researchers
• R&D in Computer, Life, Earth
and Engineering Sciences
• PhD programme, technology
transfer, public engagement
BSC-CNS is a consortium that includes:
Spanish Government 60%
Catalonian Government 30%
Univ. Politècnica de Catalunya (UPC) 10%
447 people from 44 countries *31th of December 2015

10. The MareNostrum 3 Supercomputer
Over 1015 Floating Point Operations per second
70% PRACE 24% RES 6% BSC-CNS
3 PB
of disk storage
100.8 TB
of main memory
Nearly
50,000 cores

11. Mission of BSC Scientific Departments
Earth
Sciences
CASE
Computer
Sciences
Life
Sciences
To influence the way machines are built, programmed
and used: programming models, performance tools,
Big Data, computer architecture, energy efficiency
To develop and implement global and
regional state-of-the-art models for short-
term air quality forecast and long-term
climate applications
To understand living organisms by means of
theoretical and computational methods
(molecular modeling, genomics, proteomics)
To develop scientific and engineering software to
efficiently exploit super-computing capabilities
(biomedical, geophysics, atmospheric, energy, social
and economic simulations)

12. Higgs and Englert’s Nobel for Physics 2013
Last year one of the most computer-intensive scientific
experiments ever undertaken confirmed Peter Higgs and François
Englert’s theory by making the Higgs boson – the so-called “God
particle” – in an $8bn atom smasher, the Large Hadron Collider at
Cern outside Geneva.
“ the LHC produces 1PetaByte of data every second …”

13. Pervasive
Connectivity
Explosion of
Information
400,710 ad
requests
2000 lyrics played
on Tunewiki
1,500 pings sent on
PingMe
208,333 minutes
Angry Birds played
23,148 apps
downloaded
98,000 tweets
Smart Device
Expansion
In 60 sec
today
2013
30
Billion
By 2020
40
Trillion GB
… for 8
Billion
10
Million
DATA
(1)
(2)
(3)
Devices
Mobile
Apps
(4)
(1) IDC Directions 2013: Why the Datacenter of the Future Will Leverage a Converged Infrastructure, March 2013, Matt Eastwood ; (2) & (3) IDC Predictions 2012: Competing for
2020, Document 231720, December 2011, Frank Gens; (4) http://en.wikipedia.org
A New Era of Information Technology
Current infrastructure sagging under its own weight
Internet of Things

14. The Data Deluge
2020
40ZB*
(figure exceeds prior forecasts
by 5 ZBs)
2005 2010
2012
2015
8.5ZB
2.8ZB
1.2ZB0.1ZB
* Source: IDC

15. How big is big?
10
30
This will take us
beyond our
decimal system
Geopbyte
This will be our digital
universe tomorrow…
Brontobyte
10
27
10
24This is our digital universe today
= 250 trillion of DVDs
Yottabyte
1021
1.3 ZB of network
traffic by 2016
Zettabyte
10
18
1 EB of data is created on the internet each day = 250 million DVDs
worth of information. The proposed Square Kilometer Array
telescope will generated an EB of data per day
Exabyte
10
12
Terabyte
500TB of new data per day are ingested in Facebook databases
1015
Petabyte
The CERN Large Hadron Collider
generates 1PB per second
109
Gigabyte
10
6
Megabyte
Saganbyte, Jotabyte,…

16. 16
Big Data activities at BSC
• Active research in a large
number of topics, involvement
of all departments
• Architecture & HW design for DB
acceleration
• System software
• Programming models
• Storage technologies
• Algorithms & Analytics
• Applications
• Large number of funded projects
• Training activities
• In-house projects and
applications
• Involvement in worldwide and
European initiatives
• Projects with industry
Severo Ochoa Software Stack
In-house applications Tools

17. Cognitive Computing

18. Cognitive Computing
18

19. Proyecto CaixaBank-BSC
Acuerdo de colaboración
para promover
conjuntamente el desarrollo
de sistemas avanzados de
“deep learning” con
aplicaciones a los
servicios bancarios

20. Success in H2020
BSC has the 7th highest return in Spain from H2020:
1. El Consejo Superior de Investigaciones Científicas (CSIC),
2. La Fundación Tecnalia Research and Innovation
3. Industria de Turbo Propulsores
4. Atos Spain
5. Universidad Politécnica de Madrid
6. Universidad Politécnica de Catalunya
7. Barcelona Supercomputing Center
8. La Universidad Politécnica de Valencia
9. El Centro de Regulación Genómica de Barcelona
10. Acciona Infraestructuras
Source: EFE futuro article from 17.03.2016 [link]

21. BSC & The Global Industry 2016
NVIDIA GPU
Center of Excellence
IBM-BSC
Deep Learning Center
Intel-BSC Exascale Lab
BSC-Microsoft
Research Centre

22. 22
Projects with the Energy Industry
Research into advanced technologies
for the exploration of hydrocarbons,
subterranean and subsea reserve
modelling and fluid flows
Repsol-BSC Research Center Iberdrola Renovables

23. 0
20
40
60
80
100
120
140
160
180
200
China US EU Japan Switzerland Saudi Arabia
RpeakexpressedinPFlops
June 2016 June 2015
HPC in a global competition
Top20 Supercomputers
Top 20 supercomputers, June 2016
1. China
2. China
3. United States
4. United States
5. Japan
6. United States
7. United States
8. Switzerland
9. Germany
10. Saudi Arabia
11. France
12. United States
13. Germany
14. United States
15. United States
16. United States
17. United Kingdom
18. United Kingdom
19. United States
20. United States
In 2016, top20 machines represent 40% of top500, and China is 21%, US 12% and Europe 3%
But in 2015, it was 39%, but the distribution was 10% for China, US 21%, Europe 2,6% and Japan 2,1%

24. HPC is a global competition
“The country with the strongest computing capability
will host the world’s next scientific breakthroughs”.
US House Science, Space and Technology Committee Chairman
Lamar Smith (R-TX)
“Our goal is for Europe to become one of the top 3
world leaders in high-performance computing by 2020”.
European Commission President
Jean-Claude Juncker (27 October 2015)
“Europe can develop an exascale machine with
ARM technology. Maybe we need an .
consortium for HPC and Big Data”.
Seymour Cray Award Ceremony Nov. 2015
Mateo Valero

25. HPC: a disruptive technology for Industry
“…Europe has a unique opportunity to act and
invest in the development and deployment of High
Performance Computing (HPC) technology, Big
Data and applications to ensure the
competitiveness of its research and its industries.”
Günther Oettinger, Digital Economy & Society
Commissioner
The transformational impact of excellent
science in research and innovation
Final plenary panel at ICT - Innovate, Connect,
Transform conference,
22 Oct 2015, Lisbon.

26. BSC and the EC
“"Europe needs to develop an entire
domestic exascale stack from the
processor all the way to the system and
application software",
Mateo Valero, Director of Barcelona
Supercomputing Center
Final plenary panel at ICT - Innovate,
Connect, Transform conference, 22
October 2015 Lisbon, Portugal.
the transformational impact of excellent science in research
and innovation

27. Worldwide HPC roadmaps
From Tianhe-2..
…to Tianhe-2A
with domestic
technology.
From K computer…
… to Post K
with domestic
technology.
From the PPP for
HPC…
to future PRACE
systems…
…with domestic
technology
with domestic
technology.
IPCEI on HPC
?

28. HPC Infrastructure HPC Technologies
HPC Applications
ESFRI Implementation funded 21 FET-HPC projects funded
8 centres of excellence funded
Worldwide HPC roadmaps

29. 29
Horizon 2020 and HPC
The European roadmap
Source: Panagiotis Tsarchopoulos, “The European
Supercomputing Research Programme”, November 2015

30. 30
The European H2020 FETHPC landscape
(1) Building FPGA-based prototypes (6 projects):
MANGO: heterogeneous and custom accelerators emulated in Xilinx FPGAs
Green FLASH: real-time telescope (E-ELT) controller using Altera FPGAs
EXTRA: Maxeler reconfigurable architectures
ECOSCALE: ARM + programming approach (UNILOGIC) for FPGAs
ExaNoDe: ARM + Xilinx FPGAs
ExaNEST: (ARM +) cooling and FPGA-based (prototype) interconnect
(2) Simulation-based SoC design (1 project):
Mont-Blanc 3: design of HPC SoC based on ARM and accelerators
(3) Prototyping storage (2 projects)
NEXTGenIO: prototype using Intel 3D Xpoint and Fujitsu integration
SAGE: data-centric extreme data percipient storage
(4) Software only: algorithms, mathematics, or programming models (10 projects)
AllScale, INTERTWINE, ANTAREX, NLAFET, ComPat
ExaFLOW, ExCAPE, READEX, ExaHYPE, ESCAPE

31. Mont-Blanc HPC Stack for ARM
Industrial applications
System software
Hardware
Applications

32. BSC proposal for the Exaflop
• Exaflop @ 20MW
• Target: using a 10TF node then
• Need ~100K nodes
• Each node ~200W, including memory & network
• Assume 80%:20% for cores : memory
• Using 800Mhz vector units, we need
• 100 cores, 2 v.u./core, 64 lanes/v.u., muladd/lane
• Why 2 v.u.? Assuming a 4wide ARM core driving the v.u., 2 v.u. seems
reasonable
• Hence, about 1.6W for a “core+V.U.”
• Including the ARM front end, the v.u., the caches, the interconnect slice
BSC proposal 2010

33. 33
January 2016 - IPCEI on HPC
The European roadmap
The Luxembourg government, together with France, Italy and Spain launched an “Important
Project of Common European Interest” (IPCEI) on HPC and Big Data enabled applications.
“We are convinced that Europe has a unique opportunity to act
and invest in the development and deployment of High
Performance Computing (HPC) technology, Big Data and
applications to ensure the competitiveness of its research and
its industries”.
Commissioner Oettinger, January 2016

34. Council conclusions on the "Digital Single
Market Technologies and Public Services
Modernization" package
THE COUNCIL OF THE EUROPEAN UNION:
• WELCOMES the ambitious goals towards a European Exascale
high-performance computing (HPC) capability;
• RECOGNISES the need to support the technology, infrastructure
and applications and build on EU strengths and existing initiatives
to develop a HPC ecosystem relevant for all Member States, with
the ambition of placing the EU among the top supercomputing
powers in the world by 2022 and strengthen competitiveness; and
• ACKNOWLEDGES the launch of the Important Project of Common
European Interest on HPC and Big Data enabled applications.

35. Our vision
“Building an open and trusted European HPC-BD
ecosystem for a competitive data and knowledge
economy in Europe”
The IPCEI on HPC-BD has 3 major interlinked work-packages:
Development of 2
next generation
HPC-BD exa-flop
machines
Building an open
and trusted
European HPC-BD
ecosystem
Installation of
Industry relevant
HPC-BD enabled
applications

36. BSC as IPCEI Infrastructure WP leader
Combining HPC facilities, high speed networks and Big Data
Technologies in a Cloud access model
CloudHPC Infrastructures
Big Data & Cognitive Computing
Networking

37. From Test Bed applications to
infrastructures fueled by EU technology
The planned Test Bed applications will focus on specific needs of the
participating countries
The objective is to support industries out of the participating
countries to develop new European technologies and digital services
Test Beds will be developed in the following domains:
Personalized Medicine Smart Space Smart Manufacturing Fintech
New advanced Materials Smart Agrifood Smart City Applications

38. BSC Flagship
on Personalized Medicine
Moving to
the exascale
Addressing
new challenges

39. The Example of Personalised Medicine
Data generators:
Drug
Development
Data transfer, storage, processing,
analysis and access:
Precise Medicine
Enabling…
Health Systems
Personalised
Healthcare
Pharmaceutical
Companies

40. Our current expertise
2016
Biomedical Questions
Data Management
and Primary Analysis
ICGC-PANCANCER
CLL 1
CLL 2
2016
Analysis of blood
tumors
2014 (SMUFIN)2012 20152011
Study of kidney (Rhabdoid) tumor

41. Detecting genome alterations
Chronic Lymphocytic Leukemia
We set up a Sequencing pipeline for the Chronic Lymphocytic Leukemia (CLL)
Genome Project, that aims to generate a comprehensive catalogue of genomic
alterations involved in the development and progression of the disease.
Genome
Sequencing (CNAG)
BSC
500 Patients
1 Patient = less than a day
Data Management
Around 1.5 Pb
HPC Computing
Sustained 10-15%
BSC MareNostrum-II
Experimental
Validation
Puente et al. Nature 2011
Quesada et al. Nature Genetics 2012
Kulis et al., Nature Genetics 2012
Puente et al., Nature under review

42. Selected Centers for the storage, analysis
and distribution of PanCancer data
European Bioinf.
Institute, UK
Electronics and
Telecom, Korea
IMSUT RIKEN
Japan
Barcelona Supercomputing
Center, Spain
University of
California, San Diego
University of
California, Santa Cruz
Deutsches
Krebsforschungszentrum,
Germany
The Sanger
Institute, UK
Ontario Institute of
Cancer Research, CA
• 2500 Tumor-Normal WG Pairs for
more than 20 Tumor types
• ~1500 RNAseq samples
• ~1500 Expression arrays
• ~1400 Methylation data
• Clinical Recods

43. Partnerships in
Cancer Genomics
PanCancer
Data Production
Hospitals
Partnerships with
Large Genomic
Infrastructures
ELIXIR
DATA FOR LIFE
Technology
Transfer
ICGC
Clinical Genomics

44. Next generation Human-Computer
(molecular) design tools
HPC Biophysics
Cognitive
Computing
Visualization

45. What other countries are doing?
An example to follow?

46. What other countries are doing?
An example to follow?

47. Una primera iniciativa con IBM
Centro de colaboracion nacional para la
investigacion en Medicina personalizada cognitiva

48. Necesitamos una gran alianza

49. Do we need an
type consortium for HPC and Big Data?
A window of opportunity is open:
• Basic industrial and scientific know-how is available
• Excellent funding opportunities exist in H2020 at European level and in the
member state structural funds
It’s time to invest in large Flagship projects for
HPC to gain critical mass
HPC European strategy & Innovation
http://ec.europa.eu/commission/2014-2019/oettinger/blog/mateo-valero-director-
barcelona-supercomputing-center_en

50. www.bsc.es
For further information please contact
mateo.valero@bsc.es
Muchas Gracias!!!

51. Why precision medicine?

52. Hardware/Software Codesign & Analytics for Genomics at Scale
Codesign
Software
Hardware
Advanced Analytics
Machine Learning
Deep Learning
Collaborations
DSA Interface
Workload
Orchestration
Acceleration
Workload
Orchestration
Optional
p-memory
hash table
m. buffer
m. buffer
m. buffer
m. buffer
in-memory
hash table
Input Files
Softwar
e
libbionic libflash
Big Data Architectures Neural Networks
NVM → CPU → RAM → GPU → RAM → CPU → NVM
Knowledge Discovery
Codesign: Software / Accelerators / Processors / NVM
Analysis & Visualization
Catapult:
Massive FPGA
Acceleration
Synergies

53. ICGC-TCGA PanCancer Project
Identification and
classification of
genome and
expression variation
Different Tumor types
(2000 patients, 4000 genomes)
European
Bioinformatic
Institute
University of
Chicago
Electronics and
Telecom.
Res. Inst. (Korea)
IMSUT
RIKEN
Barcelona
Supercomputing
Center
Broad Institute
MIT
Ontario institute
for cancer researchLargest world-wide
initiative in of
biomedical genomics.
Computing protocols
and resources are a
limiting factor

54. 54
Seymour Cray Prize 2015
Mateo Valero
“Europe can develop an exascale machine with ARM
technology”

55. HPC-BD in an innovation driven economy
Preparing Europe for the Digital Economy and the
Digital Single Market
– Delivering open, trusted and secure access to HPC and Big Data
facilities for industrial private companies and public actors
– Developing European exa-scale technology and accelerate HPC-BD
innovation by supporting European suppliers
– Increasing the probability innovations and new technologies will be
transformed into real-world customer services

56. Nuestro objetivo
“Actuar como catalizador nacional con relación global para proveer
al ecosistema de los avances necesarios para que, en un plazo de
10 años, la medicina personalizada cognitiva sea una realidad
desplegada en la red de servicios hospitalarios“
Secuenciación
HPC
&
Big
Data

57. BSC and the EC
“"Europe needs to develop an entire
domestic exascale stack from the
processor all the way to the system and
application software",
Mateo Valero, Director of Barcelona
Supercomputing Center
Final plenary panel at ICT - Innovate,
Connect, Transform conference, 22
October 2015 Lisbon, Portugal.
the transformational impact of excellent science in research
and innovation