2015 National Meeting and Conference

30 November – 4 December

CSIR International Convention Centre

Pretoria

PROGRAMME

Table of Contents2015 CHPC Conference Overview................................................................................................................... 2

Conference Highlights....................................................................................................................................... 3

Workshops.............................................................................................................................................. 3

Forums................................................................................................................................................... 3

Exhibition................................................................................................................................................ 3

Student Cluster Competition................................................................................................................... 3

Student Research Posters...................................................................................................................... 3

Plenary Talks.......................................................................................................................................... 3

Industrial Crossfire.................................................................................................................................. 3

2015 Tutorial Workshops and Forums..............................................................................................................4

Workshops Day 1 — Monday 30 November 2015............................................................................................5

Workshops Day 2 — Tuesday 1 December 2015.............................................................................................6

Wednesday 2 December.................................................................................................................................. 8

Thursday 3 December...................................................................................................................................... 9

Friday 4 December......................................................................................................................................... 10

Plenary Sessions............................................................................................................................................ 11

Keynote Talks................................................................................................................................................. 15

Venue Floor Plan............................................................................................................................................ 17

Map to CSIR International Convention Centre................................................................................................18

1

2015 CHPC Conference OverviewWorkshops & Forums Conference

Monday30 November

Tuesday1 December

Wednesday2 December

Thursday3 December

Friday4 December

08:00 Registration 07:30 Registration

09:00 W1 / W2 / W3 / W4 / W5 / W10 / F1 W1 / W2 / W6 / W7 / W8 / F1 08:30 OpeningPlenary 1 Giles

Plenary 5 BoultonPlenary 6 Hodson

Plenary 8 PandaPlenary 9 Ngoepe

10:30 Break 10:00 Break

11:00 W1 / W2 / W3 / W4 / W5 / W10 / F1 W1 / W2 / W6 / W7 / W8 / F1 10:45 Plenary 2 Erdody & BraamPlenary 3 Busch

4 / 5 / 6 : A7 / 8 / 9 : A

12:30 Lunch 12:15 Lunch

13:30 W1 / W2 / W3 / W4 / W5 / W10 / F1 W1 / W2 / W6 / W7 / W9 / F1 / F2 13:30 1 / 2 / 3 : A 4 / 5 / 6 : BClosingLunch

14:00

15:00 Break 15:00 Break

15:30 W1 / W2 / W3 / W4 / W5 / W10 / F1 W1 / W2 / W6 / W7 / W9 / F1 / F2 15:45 1 / 2 / 3 : BCrossfireFeldman

17:00 End of day 17:15

18:00 Plenary 4 Cornelius Plenary 7 Sterling

W1: Code Modernization [2]W2: OpenStack [1]W3: CUDA Programming [1]W4: Python [1]W5: Gaussian [1]W6: VASP [1]F1: SADC HPC Forum [2]

W7: Discovery Studio [1]W8: Infiniband [½]W9: Hadoop [½]W10: PBS Pro [1]

F2: CHPC Industry Forum [½]

1A: Mechanics1B: HPC Vendors2AB: Chemistry3AB: SA NREN

4A: Health4B: Mechanics5AB: Chemistry6AB: DIRISA

7: HPC Techniques8: Material Science9: Earth & Space

13:00

Conference Highlights

Workshops

W1. [2 days] HPC Code Modernization, Intel

W2. [1 day] OpenStack and Docker as a foundation for High Performance Computing, Praxis

W3. [1 day] Introduction to GPU Computing, Nicolin Govender, CHPC

W4. [1 day] Python — Medical Image processing, Rudolph Pienaar, BCH

W5. [1 day] Gaussian for Advanced Users, Michael Bearpark & Alexandra Simperler, Imperial College

W6. [1 day] Electronic Structure Calculations Using VASP, David Santos-Carballal, Cardiff University, and Nelson Dzade, Utrecht Universit

W7. [1 day] Docking in Discovery Studio 4.5, Stephen Pelly, Stellenbosch University

W8. [½ day] InfiniBand and High-Speed Ethernet for Dummies, DK Panda, The Ohio State University

W9. [½ day] Big Data Processing with Hadoop, DK Panda, The Ohio State University

W10. [1 day] Fundamentals of PBS Professional 13.0 as WLM Framework. [TBA], Ian Littlewood, Altair

Forums

F1. [2 days] SADC HPC Forum

F2. [½ day] CHPC Industry Forum

Exhibition

Student Cluster Competition

Student Research Posters

Plenary Talks

WEDNESDAY

P1. HPC-Enabled Innovation and Transformational Science & Engineering: The Role of Cyberinfrastructure, Merle Giles, NCSA

P2. From the South: Building Together a High-Tech Ecosystem, Nicolas Erdody, Open Parallel Ltd.

HPC Stack—Scientific Computing Meets Cloud, Peter Braam, www.peterbraam.com

P3. Research and Education in a Globalised World: Opportunities, Obstacles and a Way Forward,Rene Buch, NORDUnet

P4. The Convergence of HPC and Big Data – Challenges and Opportunities, Herbert Cornelius, Intel

THURSDAY

P5. Open Data, Big Data and the Future of Science, Geoffrey Boulton, University of Edinburgh

P6. Mobilising the data revolution: CODATA’s work on data policy, data science and capacity building, Simon Hodson, CODATA

P7. HPC at the Cross Roads – The Right Turn to Exascale, Thomas Sterling, Indiana University

FRIDAY

P8. Designing HPC and Big Data Middleware for Exascale Systems, DK Panda, The Ohio State University

P9. Simulated synthesis, characterisation and performance of nano-architectured energy storagematerials, Phuti Ngoepe, University of Limpopo

Industrial Crossfire

3

2015 Tutorial Workshops and Forums

Monday 30 November Tuesday 1 December

08:00 Registration Registration

09:00

Intel CUDA Python Gaussian PBS Pro SADC Forum Intel OpenStack VASP DiscoveryStudio Infiniband SADC Forum10:30

11:00

12:30 Lunch Lunch

13:30

Intel CUDA Python Gaussian PBS Pro SADC Forum Intel OpenStack VASP DiscoveryStudio Hadoop SADC Forum

CHPCIndustryForum

15:00

15:30

17:00 End of day 1 End of day 2

Workshops Day 1 — Monday 30 November 2015

Time: 09:00 – 17:00 Venue: Crystal

HPC Code Modernization IIntel & Bayncore

The Intel HPC Code Modernization tutorial is for Software Developers, Software Architects, Project Managers, R&D Directors and CxOs in HPC. It’s a unique opportunity to hear about future technology innovations in HPC from Intel – including next generation Intel Xeon and Intel Xeon Phi (aka Knights Landing) processors – and learn how to create, modernize and optimize high-performance parallel code in C / C++ or Fortran.

● Parallel architecture, technical advances and features of latest and future Intel processors — Xeon and Xeon Phi ● Overview of the latest Intel Compilers, MKL library and analysis tools available in Intel Parallel Studio XE ● Coding high-performance analytics in C++ using Intel’s new data Analytics acceleration library ● Optimize and perform with Intel MPI ● Best practices for vectorization –parallelism at core level (SIMD) ● Maximizing performance and scalability using performance libraries ● Tutorial: real world examples for vectorization ● Case study: dynamic load balancing of the n-body problem .

Time: 09:00 – 17:00 Venue: Diamond

OpenStack and Docker as a foundation for High Performance ComputingAslam Raffee & Lisa Herbert, Praxis; et al.

High Performance Computing (HPC) enables scientists and researchers to solve complex problems that require many computing capabilities. HPC typically utilises a message passing interface (MPI) to communicate between different nodes. Migrating HPC projects to a Cloud-based installation is very simple and does not require much additional setup.

This two part workshop will provide an overview of some of the OpenStack components that can enable and support HPC workloads, as well as provide a simple demonstration of an HPC application running on top of anOpenstack cloud.

Part one will cover the following OpenStack components and how they can assist in providing a platform for hosting your HPC applications: ● Nova Compute (KVM + Docker) ● Cinder Block Storage ● Swift Object Storage ● Neutron Networking ● OpenStack API's.

In part two we will demonstrate a real world example of an HPC application running on an OpenStack cloud. In this workshop, we will build an HPC cluster with Open MPI on an OpenStack Cloud. Next, we will run an Open MPI application on top of our cluster. By the end of this workshop, you will know how to leverage the Cloud to rapidly build and scale an HPC cluster for real-time data processing while removing the dependency on physical infrastructure.

Time: 09:00 – 17:00 Venue: Amethyst

Introduction to GPU ComputingNicolin Govender, CHPC

A hands on approach to GPU computing using C++ and Nvidia CUDA: ● overview of computer hardware ● GPU computing ● examples from Linear Algebra and Image Processing

Delegates will require a laptop and will ssh into GPU

compute nodes.

Time: 09:00 – 17:00 Venue: Ruby

Python — Medical Image processingRudolph Pienaar, Boston Children's Hospital, Harvard

The ChRIS system that has been deployed to the CHPC in April of this year under the auspices of Dr du Plessis at CUBIC allows for web-based running of medical image pipelines on the CHPC.

The “glue” to connect new pipelines to ChRIS is python, and this tutorial covers how to connect custom research pipelines to ChRIS using python. In this manner, more medical imaging processing pipelines can be easily deployed to the CHPC and used by more users.

Attendees need to be familiar with the python programming language and the Linux command line. Laptops will be required but any operating system is acceptable as flash sticks with VirtualBox and a virtual image of the development environment will be lent out during the tutorial to those not running Linux.

Time: 09:00 – 17:00 Venue: Onyx

Gaussian for Advanced UsersMichael Bearpark & Alexandra Simperler, Imperial College

This tutorial covers advanced topics with Gaussian for users: convergence issues, NMR, ONIOM and TDDFT. Theformat is a mix of lectures and hands-on practicals.

Topics: ● Convergence issues in Gaussian, lecture and discussion ● NMR lecture and hands-on ● “1H Spin-Spin Coupling in three different Isomers of Dichloroethene” ● ONIOM modelling larger molecules lecture ● Introduction to the ONIOM method hands-on ● TDDFT modelling excited states lecture ● Using TD-DFT to calculate UV-vis active transitions in Formaldehyde hands-on.

Participants need to be familiar with Gaussian and be able to perform a geometry and frequency calculation. Attendees need a laptop and ssh access to the cluster or an installation of Gaussian on their machine. We will organise a copy of GaussView.

Time: 09:00 – 17:00 Venue: Garnet

Fundamentals of PBS Professional 13.0 as WLM Framework.Ian Littlewood, Altair

• Are you interested in learning and discussing the basicsof PBS Pro?

• Are you looking at deploying PBS Pro to a computing facility in your organisation?

• Do you need to improve efficiencies by proper scheduling and queuing?

Ian Littlewood, Director of Enterprise Solutions (EMEA), Altair UK, will present an interactive workshop and discussion.

Time: 09:00 – 17:00 Venue: Emerald

SADC Forum IBy invitation only.

See separate Forum programme sheet.

5

Workshops Day 2 — Tuesday 1 December 2015

Time: 09:00 – 17:00 Venue: Crystal

HPC Code Modernization IIIntel & Bayncore

The second day of this programming tutorial is dedicated to several real world case studies illustrating the use of powerful techniques to speedup your parallel codes. These include: OpenMP, Cilk, and Thread building blocks; OpenMP and SIMD parallelization on Xeon and Xeon Phi; threading, vectorization, and enhanced data locality; thread and vector parallelism on the Xeon Phi; analyse and validate performance results using VTune.

Case studies: ● Performance optimization of Black-Scholes calculation ● Code modernization of polyhedron benchmark suite ● Improving performance of numerical weather prediction codes ● Pairwise sequence alignment with the Smith-Waterman algorithm ● Code optimization in a 3D diffusion model .

Time: 09:00 – 17:00 Venue: Diamond

OpenStack and Docker as a foundation for High Performance ComputingAslam Raffee & Lisa Herbert, Praxis; et al.

High Performance Computing (HPC) enables scientists and researchers to solve complex problems that require many computing capabilities. HPC typically utilises a message passing interface (MPI) to communicate between different nodes. Migrating HPC projects to a Cloud-based installation is very simple and does not require much additional setup.

This two part workshop will provide an overview of some of the OpenStack components that can enable and support HPC workloads, as well as provide a simple demonstration of an HPC application running on top of anOpenstack cloud.

Part one will cover the following OpenStack components and how they can assist in providing a platform for hosting your HPC applications: ● Nova Compute (KVM + Docker) ● Cinder Block Storage ● Swift Object Storage ● Neutron Networking ● OpenStack API's.

In part two we will demonstrate a real world example of an HPC application running on an OpenStack cloud. In this workshop, we will build an HPC cluster with Open MPI on an OpenStack Cloud. Next, we will run an Open MPI application on top of our cluster. By the end of this workshop, you will know how to leverage the Cloud to rapidly build and scale an HPC cluster for real-time data processing while removing the dependency on physical infrastructure.

Time: 09:00 – 17:00 Venue: Garnet

Electronic Structure Calculations Using VASPDavid Santos-Carballal, Cardiff University, and Nelson Dzade, Utrecht Universit

The Vienna Ab initio Simulation Package (VASP) is a first principles computer program for atomic scale materials modelling which provides details about the electronic structure of the systems under consideration. The workshop will cover a range of contemporary VASP applications to understanding the solid state and surface chemistries of materials. Specific topics that will be presented include the determination of equilibrium structures, electronic band structure, density of states, phonon dispersion relations, vibrational spectra, atomic charge characterisation, mechanical properties, surface relaxation, surface energy, workfunction, scanning tunnelling microscopy (STM) image, morphology of

nanocrystals, adsorption energy, redox properties of surfaces and reaction pathway search.

This one day workshop is intended for undergraduate project students, postgraduate students, post-doctoral researchers and researchers who are familiar with the fieldand want to employ state-of-the-art methodology based onthe density functional theory to understand bulk materials properties, surface science and heterogeneous catalysis phenomena.

● Introduction to Density Functional Theory ● Tutorial: Introduction to VASP Calculations for Solids and Molecules ● Hands-on exercises: ◊ Equilibrium structure determination ◊ Electronic band structure and density of states analysis ◊ Vibrational spectra calculation and phonon dispersion relations ◊ Mechanical properties ◊ Atomic charge characterization

● Tutorial: Introduction to VASP Calculations for Surfaces ● Hands-on exercises on: ◊ Surface relaxation and surface energy ◊ Workfunction ◊ STM topographic images simulation ◊ Crystal morphology ◊ Adsorption energy ◊ Local density of states (LDOS) ◊ Charge transfer characterisation ◊ Transition state identification and confirmation

● Concluding remarks and other capabilities of VASP: ◊ Magnetism ◊ Redox properties ◊ Reaction pathway search

Familiarity with UNIX or Linux environment, OpenMP and MPI is a requirement for participation in the workshop. Attendees need access to a laptop, cluster and VASP software. Origin, Excel or any open-source spreadsheet application with graphing capabilities is required for post-processing. p4vasp, VMD, HIVE, Jmol and GDIS are necessary for structure visualization, STMimages production, vibrational modes visualization and morphology construction respectively and are free software. p4vasp, HIVE, Jmol and GDIS require no installation to run.

Time: 09:00 – 17:00 Venue: Onyx

Docking in Discovery Studio 4.5Stephen Pelly, Stellenbosch University

A “hands-on” workshop on tools in Discovery Studio (DS) for receptor based drug design aimed at anybody interested in rational drug design using receptor based modelling tools available in DS. This will include protein preparation, docking, cross-docking and interpretation of results. Some ‘problem’ receptors will also be covered. Post-docking manipulations such as binding energy calculations, full receptor minimisations and rescoring willalso be looked at. The intention is also to allow enough time to cover individual requests from the participants.

● A brief introduction to the tools in DS associated with receptor based drug design ● Docking : ◊ Protein selection and preparation ◊ Ligand preparation ◊ Docking ◊ Interpretation of results ◊ Post docking tasks ● Advanced docking tasks : ◊ Handling problem receptors ◊ Waters ◊ Preparing a non-ideal receptor (this does not include homology modelling) ● Post docking visualisation manipulations ● Time permitting: an introduction to selected drug design tools (receptor based) in DS and requests from delegates.

Attendees will require a laptop with Discovery Studio 4.5installed and should have some knowledge of navigatingaround the DS interface.

6

Time: 09:00 – 12:30 Venue: Ruby

InfiniBand and High-Speed Ethernet for DummiesDK Panda, The Ohio State University

InfiniBand (IB) and High-speed Ethernet (HSE) technologies are generating a lot of excitement towards building next generation High-End Computing (HEC) systems including clusters, data centres, file systems, storage, and cloud computing (Hadoop, HBase and Memcached) environments. RDMA over Converged Enhanced Ethernet (RoCE) technology is also emerging.

This tutorial will provide an overview of these emerging technologies, their offered architectural features,their current market standing, and their suitability for designing HEC systems. It will start with a brief background behind IB and HSE. In-depth overview of the architectural features of IB and HSE (including iWARP and RoCE), their similarities and differences, and the associated protocols will be presented.

Next, an overview of the emerging OpenFabrics stack which encapsulates IB, HSE and RoCE in a unified manner will be presented. Hardware/software solutions and the market trends behind IB, HSE and RoCE will be highlighted. Finally, sample performance numbers of these technologies and protocols for different environments will be presented.

Time: 13:30 – 17:00 Venue: Ruby

Accelerating Big Data Processing with Hadoop and MemcachedDK Panda, The Ohio State University

Apache Hadoop is gaining prominence in handling Big Data and analytics. Similarly, Memcached in Web 2.0 environment is becoming important for large-scale query processing. These middleware are traditionally written with sockets and do not deliver best performance on datacenters with modern high performance networks. In this tutorial, we will provide an in-depth overview of the architecture of Hadoop components (HDFS, MapReduce, RPC, HBase, etc.) and Memcached. We will examine the challenges in re-designing the networking and I/O components of these middleware with modern interconnects, protocols (such as InfiniBand, iWARP, RoCE, and RSocket) with RDMA and storage architecture. Using the publicly available RDMA for Apache Hadoop (http://hadoop-rdma.cse.ohio-state.edu) software package, we will provide case studies of the new designs for several Hadoop components and their associated benefits. Through these case studies, we will also examine the interplay between high performance interconnects, storage systems (HDD and SSD), and multi-core platformsto achieve the best solutions for these components.

Time: 09:00 – 17:00 Venue: Emerald

SADC Forum IIBy invitation only.

See separate Forum programme sheet.

Time: 13:00 – 17:00 Venue: Amethyst

CHPC Industry ForumBy invitation only.

See separate Forum programme sheet.

7

Wednesday 2 December

07:30 Registration

08:30

VENUE: PLENARY HALL

Chair: Werner Janse Van Rensburg, CHPC

Opening and WelcomeDr Phil Mjwara, Director-General, Department of Science and Technology

Dr Happy Sithole, Director, Centre for High Performance Computing

09:00HPC-Enabled Innovation and Transformational Science & Engineering:

The Role of Cyberinfrastructure Merle Giles, NCSA

09:45 Break

10:30

VENUE: PLENARY HALLChair: Paul Calleja, University of Cambridge

From the South: Building Together a High-Tech EcosystemNicolas Erdody, Open Parallel Ltd.

HPC Stack—Scientific Computing Meets CloudPeter Braam, www.peterbraam.com

11:40Research and Education in a Globalised World: Opportunities, Obstacles and a Way Forward

Rene Buch, NORDUnet

12:30 Lunch

VENUE: CRYSTAL-GARNET-ONYX

1A: Computational MechanicsChair: Nicolin Govender, CHPC

VENUE: RUBY

2A: Computational ChemistryChair: Jeanet Conradie, UFS

VENUE: EMERALD

3A: SA NRENChair: Renier van Heerden, SANReN

13:30

KEYNOTE: Ian Littlewood, Altair: TBA KEYNOTE: Cornie van Sittert, NWU: Classification of metal carbene catalyst types for alkene metathesis: Identifying descriptors

KEYNOTE: Leon Staphorst and Duncan Greaves, SANReN and TENET: The SA NREN as enabler of the South Africa's Research Community

14:10

Oliver Oxtoby, CSIR: Improvements to the Volume of Fluid method for multiphase flow modeling

Marianne Conradie, UFS: DFT Mechanistic Investigation of a Large Molecular Catalytic System.

Andrew Wilson, Lucidview: Demonstrate how LucidView uses Big Data to add more value for Security, Trending Analysis, & Management

14:35

M. Bhamjee, UJ: The computational effect of parallel partitioning approaches based on model particle distributions within the parallel partitions of a CFD-DEM model.

Daniel Jansen van Vuuren, NWU: A Conceptual DFT Study of the Solvent Extraction of Zirconium and Hafnium CHPC Technical Helpdesk

Johann Tonsing, Netronome: Concepts in Software Defined Networking

15:00 Break

VENUE: CRYSTAL-GARNET-ONYX

1B: HPC VendorsChair: Filippo Spiga, U. Cambridge

VENUE: RUBY

2B: Computational ChemistryChair: Cornie van Sittert, NWU

VENUE: EMERALD

3B: SA NRENChair: Renier van Heerden, SANReN

15:45Mellanox: TBA KEYNOTE: Jeanet Conradie, UFS: Mono-, di- and

tetranuclear bis(β-diketonato)-dichlorotitanium (IV) complexes: Isomer distribution and hydrolysis

Erasmus Burger, UP: Communications and fieldwork in Antarctica, from a field-based scientist's point of view.

16:25Torben Kling Petersen, Seagate: HPC Data Archiving – Reality or Myth?

Stefan Coetzee, UP: Theoretical and Experimental Determination of the Magnetic Properties of SelectedOrganic-Inorganic Copper(II)Halide Hybrid Materials

Barry Irwin, Rhodes U.: Sanren as a research enabler -A darknet case study

16:50Chakravarthy Nagarajan, Intel: Intel Lustre and the future of the I/O in HPC

Martin Magu, UJ: Reaction mechanisms for the dechlorination of chlorobenzenes present in selected South Africa water systems

Roderick Mooi, SANReN: Big Data, Faster (aka Operating Innovative Networks)

17:15

18:00

VENUE: EXHIBITION HALL

The Convergence of HPC and Big Data – Challenges and OpportunitiesHerbert Cornelius, Intel

Conference DinnerStudent Poster Presentations in Foyer

Note: The venue for the refreshment breaks and lunch is the Exhibition Hall

8

15:10

Thursday 3 December07:30 Registration

08:30

VENUE: PLENARY HALL

Chair: Peter Braam

Open Data, Big Data and the Future of Science

Geoffrey Boulton, University of Edinburgh

09:15Mobilising the data revolution: CODATA’s work on data policy, data science and capacity building

Simon Hodson, CODATA

10:00 Break

VENUE: CRYSTAL-GARNET-ONYX

4A: HealthChair: Charles Crosby, CHPC

VENUE: RUBY

5A: Material ScienceChair: Alexendra Simperler, ICL

VENUE: EMERALD

6A: DIRISAChair: Wim Hugo, SAEON

10:45

KEYNOTE: Rudolph Pienaar, Boston Children's Hospital: Web 2.0 and Beyond: Leveraging Web Technologies as Middleware in Healthcare and High Performance Compute Clusters: Data, Apps, Results, Sharing and Collaboration.

10:45 Thobani Gambu, UCT: A DFT study of the interaction of atomic O and O2 with Pt nanorod edge sites: limitations in modelling the overall ORR activity

10:45 Colin Wright: NICIS and Data Intensive Research

11:20 KEYNOTE: Thomas Franz, UCT: Computational Modelling of Cardiac Infarct Therapies

11:10 Sylvia Ledwaba, U. Limpopo: Generating LiMn2O4 Nanoparticles Using Simulated Amorphization and Recrystallization Technique

11:10 Paul Laughton: Creating synergy in the South African data landscape

11:50 Malebogo Ngoepe, UCT: Incorporating biochemistry in the computational modelling of cerebral aneurysms

11:35 Cliffton Masedi, U. Limpopo: Beyond Lithium-Ion Batteries: A Computational Study on Advanced Lithium – Sulphur Battery

11:35 Martie van Deventer: Cooking with Pinotage: The data steward’s role in disseminating research data

12:10Mazin Sirry, UCT: Insights into the diastolic micromechanics of injectate-based therapy of myocardial infarction: A finite element investi gation

12:00 Moritz Braun, UNISA: Multi Domain Finite Element Calculations for Molecules in 2 and 3 Dimensions using the FEniCS framework and iso-parametr

12:00 Mark Hahnel: The Digital Academia Power Struggle

12:30 Lunch

VENUE: CRYSTAL-GARNET-ONYX

4B: Computational MechanicsChair: Oliver Oxtoby, CSIR

VENUE: RUBY

5B: Computational ChemistryChair: Michael Bearpark, ICL

VENUE: EMERALD

6B: DIRISAChair: Martie Van Deventer

13:30Quinn Reynolds, Mintek: Applications of high-performance computing in arc modelling and arc detection for direct-current smelting furnaces

KEYNOTE: Kevin J. Naidoo, UCT: Bleeding Edge to Leading Edge Methods Development in Hyper Parallel Modelling and Big Data Analytics for Chemical Biology

13:30 Wim Hugo, SAEON: Roadmap for Research Data Infrastructure: Stakeholder Requirements

13: 50: Anthony Beitz: Growing an Institution’s Research Data Management Capability

14:10Markus Erwee, Mintek: A comparison of initial 2D and 3D computational models to physical cold models for modelling oxygen lancing of pyrometal

Marina van der Merwe, UP: Molecular dynamics simulations of PIBSA-based surfactants at the supersaturated ionic interface of high internal phase 14:10 Dale Peters: Data Management for

Global Climate Change Research: a Belmont Forum e-Infrastructures & Data Management Community Strategy and Implementation Plan

Panel discussions and Q&A14:35 Penny Govender, UJ: DFT studies of Vitamin B12

analogues

15:00 Break

15:45

VENUE: PLENARY HALL

Industrial CrossfireMichael Feldman, Intersect360 Research

[ Vendors ]

17:15 Note: break to allow for judges to consult and complete selection of prize-winners.

18:30

VENUE: EXHIBITION HALL

HPC at the Cross Roads – The Right Turn to ExascaleThomas Sterling, Indiana University

Cocktails and Prize-givingPresentation of the CHPC Poster Competition Prizes

Presentation of the Fouth CHPC Student Cluster Competition Awards

9

Friday 4 December

07:30 Registration

08:30

VENUE: PLENARY HALL

Chair: Happy Sithole, CHPC

Designing HPC and Big Data Middleware for Exascale SystemsDK Panda, The Ohio State University

09:15Simulated synthesis, characterisation and performance of nano-architectured energy storage materials

Phuti Ngoepe, University of Limpopo

10:00 Break

VENUE: CRYSTAL-GARNET-ONYX

7A: HPC TechniquesChair: DK Panda, OSU

VENUE: RUBY

8A: Material ScienceChair: Phuti Ngoepe, UL

VENUE: EMERALD

9A: Earth and Space SciencesChair: Daniel Moeketsi, CHPC

10:45 KEYNOTE: Bruce Mellado, U. Wits: The Discovery of theHiggs boson and the Big Data problem

10:45 Emmanuel Igumbor, UP: Group II substitutional impurities in Ge: A density functional theory hybrid calculation

KEYNOTE: Francois Engelbrecht, CSIR: Climate modelling using quasi-uniform grids — numerical efficiency and code scalability

11:25 KEYNOTE: Filippo Spiga, U. Cambridge: Heterogeneous ARM64 and NVIDIA GPU—performance versus power.

11:05 Gandamipfa Mulatedzi, UL: Feasibility studies of TiO2 and LiTiO2 polymorphs using Density Functional based Tight Binding (DFTB+)

Nicolette Chang, CSIR: High-resolution ocean modelling towards improved carbon exchange in theCSIR Variable Resolution Earth Systems Model (VR-ESM)

11:50

11:30 Masilu Godfrey Mulaudzi, UL: Density functional theory calculation of surface properties of pyrite (100) and Depression of pyrite using TGA.

Precious Mongwe, CSIR: Understanding the Role of the Southern Ocean in Mitigating Global Climate Change: Coupled Ocean Modelling at the CHPC12:00 Extended discussion.

12:15 Torben Kling Petersen, Seagate: ExaScale HPC storage – The Next step

11:55 Mokone Roberts, NWU: Density functional theory molecular modelling and experimental particle kinetics for CO2–char gasification.

Stephanie Landman, SAWS: SAWS Research and Operational Modelling for Weather and Climate

12:40 Scott Hazelhurst , U. Wits: Using Hive and Hadoop for querying next generation sequence data: a case study

12:20 Moyahabo Chuma, U. Limpopo: Computational study of anatase TiO 2 (101) surface-supported Pd 13 nanocluster

Roelf Du Toit Strauss, NWU: Computational Heliospheric Physics

13:00

VENUE: EXHIBITION HALL

ClosingLunch

14:00 Departures

Note: The venue for the refreshment breaks and lunch is the Exhibition Hall

10

Plenary Sessions

HPC-Enabled Innovation and Transformational Science & Engineering: The Role of Cyber-infrastructure Merle Giles, NCSA

HPC-enabled innovation goes beyond discovery science and necessarily includes impact. Impact-driven science and engineering is driven by a pressing need to solve problems and provide solutions. Modern infrastructure, therefore, must match the scientific method needs of the academic community as well as the needs of applied scientists and commercial organizations. This broad approach suggests heterogeneity in data and compute systems and an ability of HPC systems to ingest and distribute large volumes of data.

NCSA is a world leader in HPC-enabled industrial innovation with deep experience in academic and commercial sectors. NCSA’s Private Sector Program (PSP) Director will speak to the common issues of serving these communities and the differentiation. The potential macro and microeconomic impact of HPC centers is significant.

Merle Giles is a tireless advocate of HPC’s role as an accelerator to industrial innovation and was recognized by HPCwire as a ‘Person to Watch’ in 2015. His NCSA team won HPCwire Editors’ Choice and Readers’ Choice awards in 2014 for collaborative technical achievements inHPC. In 2012 he received an HPCwire Editor’s Choice award for Best Government-Industry Partnership.

In recent years his Private Sector Program has partnered with nearly 60% of the manufacturers in the FORTUNE100, as well as with biotech, chemical, tech, oil and gas, and agriculture companies. He and his PSP partners were central influencers of two national digital manufacturing consortia: NDEMC (National Digital Engineering and Manufacturing Consortium) and DMDII (Digital Manufacturing and Design Innovation Institute), one of the National Network for Manufacturing Innovation (NNMI) institutes.

Giles is a co-founder of the International Industrial Supercomputing Workshop, which has HPC center members from around the world that serve industry. He is a steering committee member of IDC’s HPC User Forum and the Council onCompetitiveness HPC Advisory Committee. He is co-editor of a book published in 2015 titled ‘Industrial Applications of High-Performance Computing: Best Global Practices’, offering a global overview of HPC industrial impact with contributions from eleven countries. He earned an MBA from University of Illinois at Urbana-Champaign and undergraduate degrees in accounting and business.

From the South: Building Together a High-Tech EcosystemNicolás Erdödy, Open Parallel Ltd.

Technology empowers communities and countries. Common open platforms and international projects are generating global collaboration at scale. We are witnessingthe convergence of big data, cloud and mobile technologiesbut the continuous evolution of Open Source Software, High Performance Computing and projects like the SquareKilometre Array are triggering incredible innovations all over the world as never seen before.

This talk will present Open Parallel's contribution to the SKA project since 2013 and how this work has been inspirational to the design of TOPS (The Open Parallel Stack) — an open software platform that aims towards exascale computing. It will conclude sharing our vision on how we can leverage this new convergence between OSS, HPC and the SKA to bring prosperity for our countries — together!

Nicolás Erdödy is the Founder and CEO of OpenParallel Ltd – now working towards a platformfor exascale computing. He’s a Work Package Manager at the Square Kilometre Array project (SKA) -the largest IT development in history, leading the Software Development Environmentat the Central Signal Processor (CSP) and contributing to the Compute Platform of the Science Data Processor (SDP). He is also Director of Multicore World -a prestigious think-tank conference in Australasia.

With 35 years creating ventures in the high-tech, ICT, venture capital and education industries, Nicolás holds a Masters of Entrepreneurship from the University of Otago in New Zealand and learned Fortran at the School of Engineering of the Universidad de la República, Uruguay. He lives in Oamaru in the South Island of New Zealand, and can ask for a beer in five human languages.

HPC Stack—Scientific Computing Meets CloudPeter Braam, www.peterbraam.com

The HPC Stack is a collaborative effort between Canonical,The University of Cambridge and Parallel Scientific Inc. The aim of the project is to enhance Ubuntu Linux with private package archives that enable rapid deployment of HPC installations, leveraging Ubuntu MAAS and Juju. We will discuss the state of the current development, its usefulness to SKA and give an overview of advantages of this approach for wide area collaborative computing and the use of advanced programming environments.

Peter Braam has held senior faculty positions inMathematics and Computing at Oxford and Carnegie Mellon, prior to starting 5 startup companies (of which 4 were successful) as an entrepreneur. His best known project is the Lustre file system which powers the majority of high end HPC systems, a derivative of which is the ext4 system running on virtually all Linux systems. During the last few years, Peter has joined the research efforts at the Cavendish Laboratory in Cambridge, surrounding the SKAtelescope and future parallel computing architectures.

Creating the Network to support Advanced Services for R&E — On Open Exchange Points and Global ConnectivityRené Buch, NORDUnet

Evolution is forcing NRENs to change. In addition to beinga niche bandwidth supplier to its constituency, the NREN (National Research and Education Network) is or will very soon be required to provide the community with a multitude of services including identity, trust, system integration, and e-Infrastructure services. As the NREN evolves, the supply of bandwidth to its constituency becomes a commodity in which network infrastructure reach and a growing capacity requirement are the baseline that must be delivered at the same or lower costs.

This changes the role of the NREN from being a mere infrastructure provider to that of a solutions provider and all the way to a systems integrator bridging the gap between commercial offerings and the specific needs of theresearch and education user communities.

NRENs have the opportunity to leverage on each other's services, regionally or on a global scale where the individual NRENs will congregate in clusters and will contribute with its core competencies and services. These new types of services will have a significant impact on the traffic flow patterns and capacity requirements as the NREN infrastructure will have to be well integrated with: ● The world at large, as the university starts to offer online

11

courses, by putting a colloquium online or by starting to offer MOOCs at a serious scale not just nationally but also regionally and even globally. ● Commercial cloud providers, as that is where some of the user's services are residing. ● Other NRENs, as NRENs start to deliver services to each other to reach an economy of scale.

The new types of services call for excellent connectivity beyond what an NREN traditionally offers butit also calls for regional and global collaboration at unprecedented scale.

The challenges described require the NRENs to think differently both from a service delivery and a network infrastructure point of view. To deliver all the services required in an evolving NREN community, the NREN will have to seek alliances and find new ways of delivering services, as there simply will not be funding available for all NRENs to develop all services locally.

Senior network architects from leading NRENs around the world have been working on what is now known as the Global Network Architecture (GNA), in which open exchange points play a crucial role. This work has led to a blueprint for a next generation intercontinental interconnect for R&E Networks. Althoughthe focus of this work is on intercontinental networking, the major concepts behind the blueprint can be useful and important for NRENs all over the world.

A good example of global collaboration is the “In the Field” stories Blog (www.inthefieldstories.net) advocating the role of NRENs in the community they serve. The Blog was initiated by a group of global partners but is open for all NRENs to contribute to.

René Buch has a long background within the commercial IT and Telecommunication companies and joined NORDUnet in 2005. He has since evolved an already capable network organization to a service oriented eInfrastructure company that serves scientists, students and staff at universities and educational institutions in the Nordic countries in Europe, i.e. Norway, Sweden, Finland, Denmark, and Iceland. René heads the evolution of the NORDUnet network into a darkfiber based versatile infrastructure that enables a range of new and above-the-net services for the Nordic and global user communities. René isa member of the Global R&E Network CEO Forum, and he served 6 years as a Director for Dante Ltd.. René is a member of The Nordic Council of Ministers eScience Workgroup.

The Convergence of HPC and Big Data – Challenges and OpportunitiesHerbert Cornelius, Intel EMEA

Billions of Internet users, trillions of connected devices (the “Internet of Things”) and machine-to-machine connections on the horizon are causing a tsunami of data growth. Utilizing big data technologies one can transform data into knowledge. With this huge amount of data and the pursuit to extract all the value out of this data, we see an emerging trend for complex analytics (deep machine learning, data clustering, predictive modeling, data categorization, etc.) on massive data sets. In order to process and analyze this amount of data one need the rightamount of high performance compute and storage capabilities - thus leading to a convergence of HPC and BigData environments and infrastructures. Such converged and optimized infrastructures need to handle Petascale type of workloads where petabytes of data can be analyzed with PFLOPS of compute performance, from storing, accessing and processing to high fidelity visualization and decision making. Hence, Petascale analytics with optimized HPC infrastructures (hardware and software)

can enable new insights, better informed decisions and new discoveries across a wide range of application areas, e.g. in Life-Sciences. It also demonstrates that technical computing continues to fulfill its potential as a scientific and industrial tool for discovery and innovation.

Dr. Herbert Cornelius is Technical Director Advanced Computing at Intel EMEA. Before he was Engineering Manager in Intel’s Cluster Technologies group and Technical Marketing Manager Enterprise Computing in EMEA. He joined Intel in 1993 as Senior Computational Scientist in the Scalable Systems Division after working at Fujitsu Europe and Cray Research from 1983 to 1990. Prior he worked at the University of Karlsruhe. He received a Ph.D. degree in Mathematics and diploma degree in Mathematics and Informatics from Technical University Berlin, Germany.

Open Data, Big Data and the Future of ScienceGeoffrey Boulton, University of Edinburgh

Recent technological advances are enabling an unprecedented explosion in the generation, manipulation and dissemination of data globally. This planetary phenomenon is already creating major changes in societiesand economies, and also presents great challenges and opportunities for science. It challenges a principle, concurrent publication of concept and evidence that has been the bedrock of scientific progress in the modern era of science. But it also offers new opportunities for scientific discovery through the exploitation of so-called “big data” and this, in ways that transcends traditional disciplinary boundaries; greater involvement of a wider range of stakeholders and citizens in co-production of knowledge; and a deeper democratic engagement with the ways that scientific knowledge is created and used. Moreover, open data and open science are important issues for democracy and the future of an open society. Science must be a public and not a private enterprise that is conducted behind closed laboratory doors.

Responding to the challenges and exploiting the opportunities will depend upon new technical solutions forpresenting, sharing and analysing data; on capacity building in “data science”; and on changing the habits and norms of researchers and their institutions to create a culture of openness and data sharing.

The ‘data revolution’, and the emergence of data science as a discipline, raise important questions for international science and for the science base in Africa. The Square Kilometre Array will provide a major impetus. From a broader perspective there are major opportunities and challenges to be seized. Professor Boulton’s lecture will lay out the opportunities and challenges from an international perspective and set the scene for a discussionabout science and science policy for Africa in the age of BigData.

Professor Geoffrey Boulton is Regius Professor of Geology Emeritus at the University of Edinburgh, and former Vice-Principal of the University. He is a Fellow of the Royal Society and of the Royal Society of Edinburgh (Scotland’s national academy). He is a member of the Council of the Royal Society, chairs its Science Policy Centre, and was principal authorof its influential report Science as an Open Enterprise. Building on the recommendations ofthat report, he was responsible for creating the UK Research Data Forum and is a member of the UK Government’s Research Transparency Board.

He is the chairperson of the International Council of Science(ICSU) and President of the international Committee on Data (CODATA). His research is in the fields of environmental geology

12

and glaciology, frequently involving large and diverse data volumes, and he currently leads a major project on the Antarctic Ice Sheet. National and international awards for his science include the Kirk Bryan Award of the Geological Society of America, the Seligman Crystal of the International Glaciological Society, the Lyell Medal of the Geological Society, the Royal Medal of the Royal Geographical Society, the Croll Medal of the Quaternary Research Association, and honorary doctorates from the Universities of Heidelberg, Chalmers University, Birmingham and Keele.

Industrial CrossfireMichael Feldman, Intersect360 Research

The Industrial Crossfire pits HPC vendors against each other in a special session answering questions posed by our panel of HPC experts and the audience.

Michael Feldman is a Senior Analyst at Intersect360 Research and a 35-year veteran of the computer industry. In his role at Intersect-360 Research, Michael focuses on the company's 360-degree view market intelligence subscription service as well as its client-specific research and consulting services. Prior to joining Intersect360 Research, Michael was managing editor of HPCwire, where he gained recognition as one of the foremost opinion-makers in the HPC industry.

HPC at the Cross Roads—The Right Turn to ExascaleThomas Sterling, Indiana University

This is the pivotal year for the international HPC community planning for the unprecedented achievement of exascale computing in the first half of the next decade. Even as user institutions and system vendors are lining up to deliver and employ next generation computing systems in the performance range of 100’s of Petaflops, the longer term technology direction is being set for the delivery and more importantly the application of systems capable of two-orders of magnitude performance greater than is currently realizable. While reliance on the enabling technology roadmap is an important part of the equation for this accomplishment, it is broadly recognized that a shift in structure and methodology will be essential for full exploitation of this revolutionary new class of systems.

Further more, a greater need for national and international cooperation is being coordinated through community initiatives. One important example of this is the move to formalize and standardize deployment of runtime system software and their components. Major development of runtimes are on-going in Japan, the US, and Europe. Runtime systems will provide lightweight threads capable of employing a broader degree of parallelism to deliver the necessary concurrent performance, estimated to be on the order of billion-way simultaneous operations. An example of this is the founding of the OpenHPC consortium announced at the SC15 meeting in Austin, TX and to be managed by the Linux Foundation. So important is the future of exascale computing that the President of the United States by Executive Order created the National Strategic Computing Initiative (NSCI) to guide an integrated research and development program across federal agencies and involving industry, academia, and national laboratories. This presentation will discuss the technical advances needed to achieve exascale, highlighting advances in runtime system software and discuss the truly international thrust in this significant domain that is important to the world society.

Dr. Thomas Sterling holds the position of Professor of Intelligent Systems Engineering at the Indiana University (IU) School of Informatics and Computing as well as the Chief Scientist and Associate Director of the Center for Research in Extreme Scale Technologies (CREST). Since receiving his Ph.D from MIT in 1984 as a Hertz Fellow Dr. Sterling has engagedin applied research in fields associated with parallel computing system structures, semantics, and operation in industry, government labs, and academia.

Dr. Sterling is best known as the "father of Beowulf" for his pioneering research in commodity/Linux cluster computing. He was awarded the Gordon Bell Prize in 1997 with his collaborators for this work. He was the PI of the HTMT Project sponsored by NSF, DARPA, NSA, and NASA to explore advancedtechnologies and their implication for high-end system architectures. Other research projects included the DARPA DIVAPIM architecture project with USC-ISI, the Cray Cascade Petaflops architecture project sponsored by the DARPA HPCS Program, and the Gilgamesh high-density computing project at NASA JPL. Thomas Sterling is currently engaged in research associated with the innovative ParalleX execution model for extreme scale computing to establish the foundation principles toguide the co-design for the development of future generation Exascale computing systems by the end of this decade. ParalleX is currently the conceptual centerpiece of the XPRESS project as part of the DOE X-stack program and has been demonstrated in proof-of-concept in the HPX runtime system software. Dr. Sterling is the co-author of six books and holds six patents. He was the recipient of the 2013 Vanguard Award. In 2014, he was named a fellow of the American Association for the Advancement of Science.

Designing HPC and Big Data Middleware for Exascale Systems: Opportunities and ChallengesDhabaleswar K. (DK) Panda , The Ohio State University

This talk will focus on challenges in designing HPC and Big Data middleware for Exascale systems with millions of processors and accelerators. For the HPC domain, we will discuss about the challenges in designing runtime environments for MPI+X (PGAS- OpenSHMEM/UPC/ CAF, OpenMP, and CUDA) programming models by taking into account support for multi-/many-core, high-performance networks, GPGPUs, Intel MICs and Virtualization. Energy-aware designs and co-design schemes for such environments will also be emphasized. Features and sample performance numbers from MVAPICH2 libraries will be presented. For the Big Data domain, we will focus on high-performance and scalable designs of Hadoop (including HDFS, MapReduce, RPC, and HBase), Spark, and Memcached using native RDMA support for InfiniBand and RoCE. Finally, designs to utilize HPC clusters with Parallel File Systems (such as Lustre) for Big Data applications will be presented.

DK Panda is a Professor and University Distinguished Scholar of Computer Science and Engineering at the Ohio State University, and an IEEE Fellow. He has published over 350 papers in the area of high-end computing and networking. The MVAPICH2 (High Performance MPI and PGAS over InfiniBand, iWARP and RoCE) libraries, designed and developed by his research group (http://mvapich.cse.ohio-state.edu), are currently being used by more than 2,450 organizations worldwide (in 76 countries).

This software is empowering several InfiniBand clusters (including the 8th, 11th and 22nd ranked ones) in the TOP500 list. The RDMA packages for Apache Hadoop and Memcached together with OSU HiBD benchmarks from his group (http://hibd.cse.ohio-state.edu) are also publicly available. These libraries are currently being used by more than 130 organizations in 20 countries. More details about Prof. Panda

13

are available at http://www.cse.ohio-state.edu/~panda.

Simulated synthesis, characterisation and perfor-mance of nano-architectured energy storage materialsP.E. Ngoepe, University of Limpopo

An increasing demand for energy and a shifting to renewable energy resources, has accelerated development of energy generation and storage for use in solar energy harvesting, electric vehicles, electricity grid backups. Enhanced performances of lithium ion rechargeable batteries, such as higher capacity, improved rate capabilityand sustained capacity retention for longer cycles are now achieved with various nano-architectures.

Electronic structure and atomistic simulations are capable of elucidating structural, transport and predicting performance of battery electrodes. A major challenge associated with generating atomistic models is to capture the wealth of microstructural features observed experimentally and to follow evolvement of lithiated phases obtained after intercalation, associated with charging and discharging of batteries. One technique that is capable of spontaneously generating various nanostructures and introducing complex microstructures is the atomistic simulated amorphisation and crystallization method.

This technique, which is based on molecular dynamics, is used to illustrate simulated synthesis and characterisation of different nano-architectures, i.e. nano- spheres, sheets, rods, porous and bulk of binary metal oxides such as MnO2 and TiO2. The nano-architectures arelithiated, to imitate charging and discharging, and their structural aspects and performance are characterized by simulated X-ray diffraction patterns and voltage profiles respectively. The relationship between mechanical properties, microstructural features and electrochemical

activity in nanoporous and bulk structures is highlighted. Such connection is extended to why the ternary nano Li2MnO3, an end member of high voltage composite cathodes, is electrochemically active whilst its bulk form is inactive. Lastly, nano-architectures, associated with the Li-Mn-O ternary were synthesised from amorphous spinel nanosphere. The resulting crystallised nanostructures are characterised and a preliminary analysis reveals the presence of the layered Li2MnO3 and spinel LiMn2O4 together with a variety of defects, including grain boundaries and ion vacancies. This is a first step towards addressing the challenge of voltage fade in the composite layered spinel cathodes.

Phuti Ngoepe obtained his first degree at the University of the North (now University of Limpopo) and completed a PhD in Physics at theUniversity of Witwatersrand. He spent sabbatical at the University of Keele, UK where he commenced with work on computational modelling of materials. He is Professor of Physics at the University of Limpopo and Director of Materials Modelling Centre and holds a SARChI Chair in Computational Modelling of Materials.

His research activities range from electronic to nano scales and cover energy storage, mineral processing and alloy development. He has worked with local and international collaborators – in the UK, USA, France and Japan. He has produced several publications in this field, including chapters in two books, and presented invited papers at local and international conferences. He has supervised many honours students and up to 40 MSc and PhD students and several are currently studying under his guidance. He has also contributed significantly to science policy formulation in South Africa. He is a member of several science societies and Founder Member of the Academy of Science of South Africa. Amongst honours bestowed on him are the National Science and Technology Forum award, National Research Foundation Presidential award, National Order of Mapungubwe Silver.

14

Keynote Talks

Mono-, di- and tetranuclear bis(β-diketonato)-dichlorotitanium(IV) complexes: Isomer distributionand hydrolysisJeanet Conradie, University of the Free State

Since cancer represents one of the major causes of death inhumans worldwide, a lot of research has been conducted in order to understand carcinogenic processes and to develop new chemotherapeutic drugs. Cis-diamminedichloroplatinum(II), generally known as cis-platin, was the first metal complex that was proven to possess antitumour properties. The first non-platinum complex tested in clinical trials was cis-[Ti(ba)2(OEt)2] (Hba = benzoylacetone), also known as budotitane, which was proven to be active against a wide variety of solid tumours. However, budotitane could not progress past phase 1 clinical trials, inter alia due to the fact that it hydrolyzes extensively in water, to yield dinuclear oxygen- bridged [Ti(bzac)2O]2 which is insoluble in water. In solution, budotitane exists as an equilibrium mixture of rapidly isomerising cis isomers, however the relative cytotoxicity of these isomers has not yet been investigated.

Since the spatial arrangement of the monomeric Ti complexes and the resulting hydrolyzed species are of importance for the understanding of its anticancer activity,the purpose of this contribution is to present results of density functional theory calculations on the structure, isomer distribution, solution behaviour and partial hydrolysis of a series of Ti(β–diketonato)2Cl2 complexes. Theoretically calculated results will be used to better interpret available experimental results.

Jeanet Conradie obtained her Ph.D. in Chemistry in 2000, from the University of the Free State, South Africa, where she now is a Professor in Physical Chemistry. Her research interests involve the synthesis, computational chemistry, electrochemistry and kinetics of novel transition metal complexes and their intermediates. She applies computational methods to predict or shed light on experimental observation.

In 2014, Prof Conradie was runner-up in the senior category for Distinguished Women Researchers: Physical and Engineering Science in the Department of Science and Technology’s 2014 Women in Science Awards. She also received the Merck Medal for 2014 from the South African Chemistry Institute (SACI), for her paper Redox potentials of ligands and complexes. A DFT approach, S. Afr. J. Chem. 2011, 64, 203-209. She is a member of the Academy of Science of South Africa (ASSAf, C1 NRF rated with an H-index of 23. Up to date she has 133 academic publications and 107 conference contributions in her field of research, and 13 MSc and PhD students graduated under her study guidance. She also acts as an editorial board member of the South African Journal of Chemistry, regularly serves as a referee for international/ISI/IBSS journals and, by invitation, delivered several keynote addresses at conferences. She has productive research networks, both local and international.

Computational Modelling of Cardiac Infarct TherapiesThomas Franz, University of Cape Town

Cardiovascular diseases (CVD) including myocardial infarction (MI) are the leading cause of death globally withsubstantial increases predicted by the World Health Organisation in the coming decades.

There has been significant interest in biomaterial delivery as a therapeutic approach for myocardial infarction. Computational modelling has begun to be used to improve the understanding of the biomechanical interactions between biomaterial and infarcted

myocardium to advance therapeutic biomaterial injections.

This presentation provides an overview of our latest research activities in cardiac mechanics with emphasis on intramyocardial biomaterial injections for the treatment ofmyocardial infarction. This will include aspects of solid and fluid mechanics spanning from the delivery of the biomaterials to therapeutic effects of the injectates in situ.

Thomas Franz is Associate Professor and Head of the Division of Biomedical Engineering in the Department of Human Biology, Faculty of Health Sciences at the University of Cape Town.His research focusses on therapies for myocardial infarction and prevention of heart failure, and cellular and viral mechanobiology in non-communicable and infectious diseases. Thomas’ research group comprises two postdoctoral fellows, four PhD students and three Master students.

Web 2.0 and Beyond: Leveraging Web Technologies as Middleware in Healthcare and High Performance Compute Clusters — Data, Apps, Results, Sharing and CollaborationRudolph Pienaar, Boston Children's Hospital, Harvard

Traditionally, large healthcare institutions (i.e. hospitals) can be slow to adopt new information-based technologies. This is due to many factors: partially the fact that clinical workflows change slowly and are subject to regulatory oversight; partially because historically from an information perspective healthcare has favored the rise of disconnected information silos; and partially since the actual current needs of most users in healthcare information-based technologies is relatively straight forward. However, as we approach the third decade of the 21st century, the role of value-added information and its importance in a healthcare setting is projected to increase by several orders of magnitude. Downstream processing, where information collected by a device is rapidly analyzedby sophisticated workflows running on clusters, will become more and more important. And finally, real time sharing and collaboration between interconnected teams will become the norm.

The most apt model we currently have for managing large scale information, processing, and sharing, is the Web.

This talk will explore various ways in which Web-based technologies are currently used in healthcare, and where they will need to adapt and be re-tooled over the coming years as information processing becomes ever more critical, as well as the role of HPC in healthcare.

Finally, this talk will discuss the ChRIS system developed at Boston Children's Hospital and also deployedto the CHPC and show how it is a nascent web-based middleware platform that is designed as a framework to provide unified interface to data, HPC, and collaboration.

Rudolph is the Technical Director of the Fetal-Neonatal Neuroimaging and Developmental Science Center at Boston's Children's Hospital. In this capacity he leads several informatics-based projects geared at simplifying access to data and allowing users to process this data in complex workflows, especially on clusters and cloud-based systems. He is also responsible for maintaining a dedicated in-house compute cluster for radiological image research, as well as the larger informatics needs of the Developmental Science Center.

His other research interests include characterizing brain surface features using curvature-based metrics and using statistical approaches to classify normal and abnormal development. He is also interested in white matter tractography

15

and connectivity, especially from an information-flow perspective."

Bleeding Edge to Leading Edge Methods Development in Hyper Parallel Modelling and Big Data Analytics for Chemical Biology Kevin J. Naidoo, University of Cape Town

Using Big Data Analytics methods on gene expression datawe are able to track enzymes that are key to the biochemical networks that define tumour growth in Cancer. I will relate how informatics methods are central to identifying enzymes that are targets for treatment therapies. However, once the enzymes have been discovered highly accurate modelling of the reactions they catalysed is necessary to design next generation therapies for cancer. Here the simulations of enzyme catalysed chemical reactions depend on thorough sampling and accurate models. Previously we have generalized our adaptive reaction coordinate force biasing method to perform calculations on multidimensional reaction coordinates. We termed this the Free Energies from Adaptive Reaction Coordinate Forces (FEARCF) method. This has largely solved the problem of reliable reaction space sampling for complex chemical systems. This is not sufficient though as an high level of theory that is computationally accessible is needed to accurately compute the electronic and conformational nature of transition states. To solve this problem we developed a Quantum Supercharger Library (QSL) that can be “plugged into” legacy codes such as GAMESS-UK, GAMESS-US, NWChem, etc. In one of its modules the QSL accelerates the complete Hartree-Fock computation on Graphical Processing Units. A second module of the QSL accelerates the one and two-electron derivatives that amongst others greatly speeds up optimization calculations and QM/MM simulations when plugged into commonly used ab initio electronic structure engines such as GAMESS-UK. I will discuss the development of informatics and computational methods and the necessity to link them to make meaningful discoveries in chemistry, chemical biology, molecular biology as well as material science.

Professor Kevin J. Naidoo is the director of the Scientific Computing Research Unit and a SARChI chair in Scientific Computing. He led the scientific effort to establish the Centre for High Performance Computing (CHPC) and served as acting research Director in 2006–2007. He is strongly supported by hardware giant Nvidia Corporation in its professor partnership programme as well as being a member of its international development team for the life sciences. He is a chemist by training, that has made an exceptional impact in the broader interdisciplinary field of scientific computing – constructing mathematical modelsand quantitative analysis techniques to analyze and solve scientific problems in a number of domains.

Kevin develops computer algorithms that allow scientists to

study chemical reactions leading to the structure and function of complex biological molecules that can be traced back to genomic patterns. His work has been featured on the front covers of the leading computational chemistry journals and has been published several times in the authoritative Journal of the American Chemical Society. His development of complex IT methods is motivated by the finding of solutions to pressing scientific problems particularly the development of a molecular understanding of cancer. He has chaired UNIDO workshop on rational design and a WATOC conference on computational and informatics methods and applications in chemistry and biology. He is a member of the editorial boards of major journals in computational chemistry and carbohydrate chemistry (glycobiology). Kevin Naidoo’s approach to linking biology and chemistry through tracing of the origins of biochemical pathways that changes during the onset of cancer. This pursuit led the Naidoo group to a discovery of a molecular classification method for cancer type in a patient from the gene expression profiles of a small subset of the human genome.

The Discovery of the Higgs boson and the Big Data problemBruce Mellado, University of the Witwatersrand

The discovery of the Higgs boson by experiments at the Large Hadron Collider is strongly connected with the Big Data problem. The Large Hadron Collider collides clouds of protons at a rate of 40 MHz. Every time two clouds of protons go through each other tens of proton-proton collisions take place. This leads to data flows of the order of petabytes per second and experiment. Fast real-time decisions are made to strongly reduce data flows that can be shipped to long-term storage for off-line processing. Data in storage has reached the Exabyte scale. Implemented solutions for these challenges in real-time data processing, distributed and cloud computing will be reviewed. Prospects for the upgrade of the Large Hadron Collider in the 20s will also be discussed.

A PhD from Columbia University, Bruce Mellado, a Personal Professor at the University of the Witwatersrand, is the group leader of the High-energy physics group and the director of the high-throughput electronics lab at the School of Physics. Prof. Mellado is an expert on the Higgs boson – a sub-atomic particle that is thought to give matter its mass – and was a leading participant in its discovery that was announced in 2012 and led to the Nobel Prize in Physics being awarded in 2013 to François Englert and Peter W. Higgs. He has attained and continue to have a number of positions of leadership internationally and organise conferences and workshops.

The upgrade of the Large Hadron Collider and the study of the Higgs boson particle require a magnitude increase of data flows. It is for this reason that Prof. Mellado is passionate about the Massive Affordable Computing (MAC) project. The MAC project is connected to the research needs of particle physics and that is to find a technological way to solve the problem of Big Data. We have to do that in a way that is affordable to the country and that can generate spin-offs that will help us to develop cheap computers for South Africa.

16

Venue Floor Plan

17

FRONT ENTRANCE

Map to CSIR International Convention Centre

http://www.csiricc.co.za/map/

18