Welcome to the 12th CHPC National Conference, to be held at the Century City Convention Centre, Cape Town.
The theme for the conference is: Transforming the Future through HPC and Transforming HPC for the Future
The comprehensive programme will include national and international contributions as well as contributions from our colleagues in the cyberinfrastructure system: the South African National Research Network (SANReN) and the Data Intensive Research Initiative of South Africa (DIRISA). The 2nd and 6th December will be tutorials and workshops, while the main conference will be 3–5 December. Once again, the South African Development Community (SADC) HPC Collaboration Forum will form part of the conference to discuss the SADC HPC framework and implementation plans.
Our expo zone will showcase solutions by leading technology companies and the student competition battleground where 20 teams from universities across the country will be seeking national honours in the 7th Student Cluster Competition and in the 2nd Student Cyber-Security Competition. We trust you will find an exciting programme and we look forward to meeting our regular and new delegates.
| Morning | Afternoon | Evening | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Sun | Workshops | Workshops | |||||||
| Mon | Opening & Keynote | Breakaways | Breakaways | Vendor Crossfire | Keynote | Banquet | |||
| Tue | Keynotes | Breakaways | Breakaways | Women in HPC BoF | |||||
| Wed | Keynotes | Breakaways | Breakaways | CHPC Users BoF | Keynote | Prize-giving | |||
| Thu | Workshops | Workshops | |||||||
THERE WILL BE NO DISCOUNTED REGISTRATIONS AFTER 29 NOVEMBER 2018. WALK-IN REGISTRATIONS ARE FULL PRICE.
Description:
MedeA-UNiversal CLuster Expansion (UNCLE) expands access to materials and properties at the meso and micro scales. Performing VASP ab-initio calculations on automatically chosen sets of small models, MedeA-UNCLE captures the configurational complexity of real materials at different temperatures by means of Monte Carlo random sampling. Tight integration with job control in MedeA guarantees stability and fault tolerance. Graphical tools monitor progress of fully automated simulations and allow a ready-for-use visualization of results.
Key Benefits of MedeA-UNCLE:
• Models systems containing millions of atoms with DFT accuracy
• User-friendly setup within MedeA Environment
• Workflow-based automation of cluster expansion refinement
• Efficient handling of hundreds of input structures
• Intuitive graphical analysis and visualization
• Split and restart complex calculations
• Extend and expand existing Cluster Expansions
Computational Characteristics:
• Use Genetic Algorithm or Compressive Sensing
• Full integration with MedeA-VASP and other modules
• High throughput using the JobServer
Properties from MedeA-UNCLE:
• Structures of stable phases
• Vacancy concentrations
• Miscibility
• Random mixing energy
• Phase stability as a function of temperature and concentration
• Solubility
• Order-disorder transition temperature
• Micro structure
Target Audience: This one day workshop is intended for undergraduate project students, postgraduate students, postdoctoral researchers and researchers who are familiar with the field and want to employ state-of-the art methodology based on the density functional theory to understand bulk materials properties, phase diagrams, order- disorder transition temperature, vancacy concentrations, miscibility and micro structure.
Type of tutorial: Mix of tutorials and hands-on (mostly practical)
Special requirements:
• Attendees need access to a laptop or workstation, active CHPC user account cluster.
• Open source spreadsheet such as Microsoft Excel with graphing capabilities is required for post-processing of output data.
Outline of full syllabus:
1. Introduction to the MedeA® Software environment
a. Structure retrieval, building
b. Flowchart interface
c. High Throughput support
2. Introduction to Cluster Expansion techniques
a. Theoretical background
b. MedeA® Uncle overview
3. MedeA® UNCLE hands-on (bulk)
a. Selected tutorials
4. MedeA® UNCLE hands-on (surfaces)
a. Selected tutorials
5. Q&A, wrap-up
Description:
The goal of this workshop is to introduce students to CUDA and provide them with an understanding of parallel programming. CUDA is more than a few new keywords. One must understand SIMD and the pitfalls of serialization.
Students will leave with basic CUDA skills and some OpenACC knowledge plus useful machine learning and big data tools as well. My “from Hello World to exascale machine learning in one slide” will also be covered as data parallel training fits on GPUs nicely.
Target Audience:
Anyone with C/C++ programming skills in the Unix environment who wishes to learn about parallel programming and CUDA. The material will be 60% beginner, 30% intermediate, and 10% advanced.
Prerequisites:
C/C++ along with an ability to edit and compile programs in a Unix environment
Special requirements:
Users will have access to a CHPC system with GPUs.
Attendees should bring their own laptops. The ability to view pdf or PowerPoint files is required.
Outline of full syllabus:
08:00 Registration
09:00 Introduction and morning talk (30 minutes)
Login details and extracting the workshop material (15 minutes)
Section 01: Parallel intro and a first CUDA program)
Section 02: Profiling on a GPU
10:30 Morning Refreshment Break
11:00 Section 03: More CUDA and the Thrust Interface
Section 04: “From Hello World to TF/s machine learning”
12:30 Lunch
13:30 Afternoon talk
Section 05: Controlling parallel resources
Section 06: C++ objects and transparent host/GPU data movement
15:00 Afternoon Refreshment Break
15:30 Section 07: Task level parallelism on a GPU
Section 08: Managing big data, CUDA as a scripting language via dynamic load/link
17:00 End of Day
Additional Comments:
Students can work at their own pace.
Introductory students will learn the basics of CUDA and the profiler as well as how to think in parallel and understand the impact of parallel hardware on performance.
Intermediate/advanced students will hone their thinking about parallel programming and the limitations and advantages of GPU hardware. Extra credit exercises will challenge them.
All students will learn how to use machine learning and the ability to explore this hot field and leave with a tool that allows them to train and predict using their own data sets and neural network architectures. Further, they learn how to work with and collaborate using big data.
Where to begin? HPC as a field, a technology, and a tool for a diversity of disciplines that are enabled by it is an interplay of foundational concepts, rapidly evolving knowledge, and skill sets including programming. This one-day tutorial is presented as a beginners’ guide to supercomputing to expose the novice to the breadth of issues needed to begin to understand and use such systems for real-world problems. Included are basic concepts, HPC architecture, benchmarking, parallel programming using OpenMP and MPI, basics of GPU accelerators, and libraries. Participants will be shown live demonstrations of all of these techniques and given opportunities for hands-on experience using the CHPC Petaflops scale supercomputer in South Africa. Questions will be welcome from the attendees throughout the presentations.
Schrödinger is a leading provider of scientific software in the drug design industry. The workshop will give an overview of the drug discovery tools available from Schrödinger.
Participants will be introduced to our graphical interface, Maestro 11, and get some hands-on experience with common tasks like docking and pharmacophore based virtual screening. We will also cover the newly redesigned enumeration and virtual library design tools available.
The workshop will finish with a lecture on more advanced modeling tools, and an open Q&A session.
CHPC Cyber Security Symposium (CCSS) – 1 full day
Topic: Understanding Risk in Shared CyberEcosystems workshop
As Technology continues to evolve so are the opportunities and challenges it provides. As a Society that relies largely on technology to bring us benefits, it also exposes its users to threats by the very nature of the opportunities it presents, thus becoming a focal point for Cybercrime, industrial espionage, and Cyberattacks. Protecting against these threats is of paramount priority.
Cyber Security domain is inherently dynamic, not only does system configuration changes frequently with new releases and patches, but also new attacks and vulnerabilities are regularly discovered. The core threat in Cyber Security is human, hence intelligent in nature. The attacker adapts to the situation, target environment, and countermeasures.
Attacks actions are driven by attacker’s exploratory nature, thought process, motivation, strategy, and preferences.
The goal of this workshop is to address some of the challenges faced by Network and Security Administrators in their institutions.
The workshop will address challenges such as:
• Iot Threats
• Blockchain Revolution
• Ransomware Evolution
• Future of Cybersecurity
• New hacking Methodologies
• Phishing
• Distributed Denial of Service (DDoS)
• Malware
• Internal Privilege Misuse
• Threats actors
• How to protect your environment
The following are our invited speakers:
• Prof Elmarie Bierman: Directory of SA Cyber Security Institute
• Dr Noelle Cowling : Stellenbosch University
• Dr Jabu Mtswene : CSIR
• Dr Noluxolo Gcaza: CSIR
• Mr Justine Westcott
• Mr Sikhumbuzo Mthombeni: Demension Data
• Mr Keti Cedric : Demension Data
• Mr Richard Hlalele: University of Johannesburg, Senior Manager ICT Strategy and Governance
The following are workshop Coordinators:
Mr Bigani Sehurutshi: University of Botswana
Mrs Lee-Anne Benjamin: Manager ,Corporate IT Compliance NRF
Description:
MedeA-UNiversal CLuster Expansion (UNCLE) expands access to materials and properties at the meso and micro scales. Performing VASP ab-initio calculations on automatically chosen sets of small models, MedeA-UNCLE captures the configurational complexity of real materials at different temperatures by means of Monte Carlo random sampling. Tight integration with job control in MedeA guarantees stability and fault tolerance. Graphical tools monitor progress of fully automated simulations and allow a ready-for-use visualization of results.
Key Benefits of MedeA-UNCLE:
• Models systems containing millions of atoms with DFT accuracy
• User-friendly setup within MedeA Environment
• Workflow-based automation of cluster expansion refinement
• Efficient handling of hundreds of input structures
• Intuitive graphical analysis and visualization
• Split and restart complex calculations
• Extend and expand existing Cluster Expansions
Computational Characteristics:
• Use Genetic Algorithm or Compressive Sensing
• Full integration with MedeA-VASP and other modules
• High throughput using the JobServer
Properties from MedeA-UNCLE:
• Structures of stable phases
• Vacancy concentrations
• Miscibility
• Random mixing energy
• Phase stability as a function of temperature and concentration
• Solubility
• Order-disorder transition temperature
• Micro structure
Target Audience: This one day workshop is intended for undergraduate project students, postgraduate students, postdoctoral researchers and researchers who are familiar with the field and want to employ state-of-the art methodology based on the density functional theory to understand bulk materials properties, phase diagrams, order- disorder transition temperature, vancacy concentrations, miscibility and micro structure.
Type of tutorial: Mix of tutorials and hands-on (mostly practical)
Special requirements:
• Attendees need access to a laptop or workstation, active CHPC user account cluster.
• Open source spreadsheet such as Microsoft Excel with graphing capabilities is required for post-processing of output data.
Outline of full syllabus:
1. Introduction to the MedeA® Software environment
a. Structure retrieval, building
b. Flowchart interface
c. High Throughput support
2. Introduction to Cluster Expansion techniques
a. Theoretical background
b. MedeA® Uncle overview
3. MedeA® UNCLE hands-on (bulk)
a. Selected tutorials
4. MedeA® UNCLE hands-on (surfaces)
a. Selected tutorials
5. Q&A, wrap-up
Description:
The goal of this workshop is to introduce students to CUDA and provide them with an understanding of parallel programming. CUDA is more than a few new keywords. One must understand SIMD and the pitfalls of serialization.
Students will leave with basic CUDA skills and some OpenACC knowledge plus useful machine learning and big data tools as well. My “from Hello World to exascale machine learning in one slide” will also be covered as data parallel training fits on GPUs nicely.
Target Audience:
Anyone with C/C++ programming skills in the Unix environment who wishes to learn about parallel programming and CUDA. The material will be 60% beginner, 30% intermediate, and 10% advanced.
Prerequisites:
C/C++ along with an ability to edit and compile programs in a Unix environment
Special requirements:
Users will have access to a CHPC system with GPUs.
Attendees should bring their own laptops. The ability to view pdf or PowerPoint files is required.
Outline of full syllabus:
08:00 Registration
09:00 Introduction and morning talk (30 minutes)
Login details and extracting the workshop material (15 minutes)
Section 01: Parallel intro and a first CUDA program)
Section 02: Profiling on a GPU
10:30 Morning Refreshment Break
11:00 Section 03: More CUDA and the Thrust Interface
Section 04: “From Hello World to TF/s machine learning”
12:30 Lunch
13:30 Afternoon talk
Section 05: Controlling parallel resources
Section 06: C++ objects and transparent host/GPU data movement
15:00 Afternoon Refreshment Break
15:30 Section 07: Task level parallelism on a GPU
Section 08: Managing big data, CUDA as a scripting language via dynamic load/link
17:00 End of Day
Additional Comments:
Students can work at their own pace.
Introductory students will learn the basics of CUDA and the profiler as well as how to think in parallel and understand the impact of parallel hardware on performance.
Intermediate/advanced students will hone their thinking about parallel programming and the limitations and advantages of GPU hardware. Extra credit exercises will challenge them.
All students will learn how to use machine learning and the ability to explore this hot field and leave with a tool that allows them to train and predict using their own data sets and neural network architectures. Further, they learn how to work with and collaborate using big data.
Where to begin? HPC as a field, a technology, and a tool for a diversity of disciplines that are enabled by it is an interplay of foundational concepts, rapidly evolving knowledge, and skill sets including programming. This one-day tutorial is presented as a beginners’ guide to supercomputing to expose the novice to the breadth of issues needed to begin to understand and use such systems for real-world problems. Included are basic concepts, HPC architecture, benchmarking, parallel programming using OpenMP and MPI, basics of GPU accelerators, and libraries. Participants will be shown live demonstrations of all of these techniques and given opportunities for hands-on experience using the CHPC Petaflops scale supercomputer in South Africa. Questions will be welcome from the attendees throughout the presentations.
CHPC Cyber Security Symposium (CCSS) – 1 full day
Topic: Understanding Risk in Shared CyberEcosystems workshop
As Technology continues to evolve so are the opportunities and challenges it provides. As a Society that relies largely on technology to bring us benefits, it also exposes its users to threats by the very nature of the opportunities it presents, thus becoming a focal point for Cybercrime, industrial espionage, and Cyberattacks. Protecting against these threats is of paramount priority.
Cyber Security domain is inherently dynamic, not only does system configuration changes frequently with new releases and patches, but also new attacks and vulnerabilities are regularly discovered. The core threat in Cyber Security is human, hence intelligent in nature. The attacker adapts to the situation, target environment, and countermeasures.
Attacks actions are driven by attacker’s exploratory nature, thought process, motivation, strategy, and preferences.
The goal of this workshop is to address some of the challenges faced by Network and Security Administrators in their institutions.
The workshop will address challenges such as:
• Iot Threats
• Blockchain Revolution
• Ransomware Evolution
• Future of Cybersecurity
• New hacking Methodologies
• Phishing
• Distributed Denial of Service (DDoS)
• Malware
• Internal Privilege Misuse
• Threats actors
• How to protect your environment
The following are our invited speakers:
• Prof Elmarie Bierman: Directory of SA Cyber Security Institute
• Dr Noelle Cowling : Stellenbosch University
• Dr Jabu Mtswene : CSIR
• Dr Noluxolo Gcaza: CSIR
• Mr Justine Westcott
• Mr Sikhumbuzo Mthombeni: Demension Data
• Mr Keti Cedric : Demension Data
• Mr Richard Hlalele: University of Johannesburg, Senior Manager ICT Strategy and Governance
The following are workshop Coordinators:
Mr Bigani Sehurutshi: University of Botswana
Mrs Lee-Anne Benjamin: Manager ,Corporate IT Compliance NRF
Description:
MedeA-UNiversal CLuster Expansion (UNCLE) expands access to materials and properties at the meso and micro scales. Performing VASP ab-initio calculations on automatically chosen sets of small models, MedeA-UNCLE captures the configurational complexity of real materials at different temperatures by means of Monte Carlo random sampling. Tight integration with job control in MedeA guarantees stability and fault tolerance. Graphical tools monitor progress of fully automated simulations and allow a ready-for-use visualization of results.
Key Benefits of MedeA-UNCLE:
• Models systems containing millions of atoms with DFT accuracy
• User-friendly setup within MedeA Environment
• Workflow-based automation of cluster expansion refinement
• Efficient handling of hundreds of input structures
• Intuitive graphical analysis and visualization
• Split and restart complex calculations
• Extend and expand existing Cluster Expansions
Computational Characteristics:
• Use Genetic Algorithm or Compressive Sensing
• Full integration with MedeA-VASP and other modules
• High throughput using the JobServer
Properties from MedeA-UNCLE:
• Structures of stable phases
• Vacancy concentrations
• Miscibility
• Random mixing energy
• Phase stability as a function of temperature and concentration
• Solubility
• Order-disorder transition temperature
• Micro structure
Target Audience: This one day workshop is intended for undergraduate project students, postgraduate students, postdoctoral researchers and researchers who are familiar with the field and want to employ state-of-the art methodology based on the density functional theory to understand bulk materials properties, phase diagrams, order- disorder transition temperature, vancacy concentrations, miscibility and micro structure.
Type of tutorial: Mix of tutorials and hands-on (mostly practical)
Special requirements:
• Attendees need access to a laptop or workstation, active CHPC user account cluster.
• Open source spreadsheet such as Microsoft Excel with graphing capabilities is required for post-processing of output data.
Outline of full syllabus:
1. Introduction to the MedeA® Software environment
a. Structure retrieval, building
b. Flowchart interface
c. High Throughput support
2. Introduction to Cluster Expansion techniques
a. Theoretical background
b. MedeA® Uncle overview
3. MedeA® UNCLE hands-on (bulk)
a. Selected tutorials
4. MedeA® UNCLE hands-on (surfaces)
a. Selected tutorials
5. Q&A, wrap-up
Description:
The goal of this workshop is to introduce students to CUDA and provide them with an understanding of parallel programming. CUDA is more than a few new keywords. One must understand SIMD and the pitfalls of serialization.
Students will leave with basic CUDA skills and some OpenACC knowledge plus useful machine learning and big data tools as well. My “from Hello World to exascale machine learning in one slide” will also be covered as data parallel training fits on GPUs nicely.
Target Audience:
Anyone with C/C++ programming skills in the Unix environment who wishes to learn about parallel programming and CUDA. The material will be 60% beginner, 30% intermediate, and 10% advanced.
Prerequisites:
C/C++ along with an ability to edit and compile programs in a Unix environment
Special requirements:
Users will have access to a CHPC system with GPUs.
Attendees should bring their own laptops. The ability to view pdf or PowerPoint files is required.
Outline of full syllabus:
08:00 Registration
09:00 Introduction and morning talk (30 minutes)
Login details and extracting the workshop material (15 minutes)
Section 01: Parallel intro and a first CUDA program)
Section 02: Profiling on a GPU
10:30 Morning Refreshment Break
11:00 Section 03: More CUDA and the Thrust Interface
Section 04: “From Hello World to TF/s machine learning”
12:30 Lunch
13:30 Afternoon talk
Section 05: Controlling parallel resources
Section 06: C++ objects and transparent host/GPU data movement
15:00 Afternoon Refreshment Break
15:30 Section 07: Task level parallelism on a GPU
Section 08: Managing big data, CUDA as a scripting language via dynamic load/link
17:00 End of Day
Additional Comments:
Students can work at their own pace.
Introductory students will learn the basics of CUDA and the profiler as well as how to think in parallel and understand the impact of parallel hardware on performance.
Intermediate/advanced students will hone their thinking about parallel programming and the limitations and advantages of GPU hardware. Extra credit exercises will challenge them.
All students will learn how to use machine learning and the ability to explore this hot field and leave with a tool that allows them to train and predict using their own data sets and neural network architectures. Further, they learn how to work with and collaborate using big data.
Where to begin? HPC as a field, a technology, and a tool for a diversity of disciplines that are enabled by it is an interplay of foundational concepts, rapidly evolving knowledge, and skill sets including programming. This one-day tutorial is presented as a beginners’ guide to supercomputing to expose the novice to the breadth of issues needed to begin to understand and use such systems for real-world problems. Included are basic concepts, HPC architecture, benchmarking, parallel programming using OpenMP and MPI, basics of GPU accelerators, and libraries. Participants will be shown live demonstrations of all of these techniques and given opportunities for hands-on experience using the CHPC Petaflops scale supercomputer in South Africa. Questions will be welcome from the attendees throughout the presentations.
Description:
This session will cover various aspects of analysing and optimising parallel programs to achieve optimum performance on the supercomputer. Tuning and Analysis Utilities (TAU) callgraph visualization system will demonstrate how to analyse different objects (e.g. modules, routines and functions) and identify performance bottlenecks within parallel applications.
Target Audience:
Parallel program users, developers and administrators
Prerequisites:
Attendees must at least have knowledge of compiling and running parallel applications on a cluster.
Type of Tutorial:
Mix of lectures and practicals
Special Requirements:
No need for attendees to have laptops. Presenter will demonstrate practical exercises on his laptop.
Outline:
08:00
Registration
09:00
10:30
Morning Refreshment Break
11:00
12:30
Lunch
13:30
Tuning and optimisation of parallel programs on the CHPC supercomputer
15:00
Afternoon Refreshment Break
15:30
Tuning and optimisation of parallel programs on the CHPC supercomputer
17:00
End of Day
CHPC Cyber Security Symposium (CCSS) – 1 full day
Topic: Understanding Risk in Shared CyberEcosystems workshop
As Technology continues to evolve so are the opportunities and challenges it provides. As a Society that relies largely on technology to bring us benefits, it also exposes its users to threats by the very nature of the opportunities it presents, thus becoming a focal point for Cybercrime, industrial espionage, and Cyberattacks. Protecting against these threats is of paramount priority.
Cyber Security domain is inherently dynamic, not only does system configuration changes frequently with new releases and patches, but also new attacks and vulnerabilities are regularly discovered. The core threat in Cyber Security is human, hence intelligent in nature. The attacker adapts to the situation, target environment, and countermeasures.
Attacks actions are driven by attacker’s exploratory nature, thought process, motivation, strategy, and preferences.
The goal of this workshop is to address some of the challenges faced by Network and Security Administrators in their institutions.
The workshop will address challenges such as:
• Iot Threats
• Blockchain Revolution
• Ransomware Evolution
• Future of Cybersecurity
• New hacking Methodologies
• Phishing
• Distributed Denial of Service (DDoS)
• Malware
• Internal Privilege Misuse
• Threats actors
• How to protect your environment
The following are our invited speakers:
• Prof Elmarie Bierman: Directory of SA Cyber Security Institute
• Dr Noelle Cowling : Stellenbosch University
• Dr Jabu Mtswene : CSIR
• Dr Noluxolo Gcaza: CSIR
• Mr Justine Westcott
• Mr Sikhumbuzo Mthombeni: Demension Data
• Mr Keti Cedric : Demension Data
• Mr Richard Hlalele: University of Johannesburg, Senior Manager ICT Strategy and Governance
The following are workshop Coordinators:
Mr Bigani Sehurutshi: University of Botswana
Mrs Lee-Anne Benjamin: Manager ,Corporate IT Compliance NRF
Description:
MedeA-UNiversal CLuster Expansion (UNCLE) expands access to materials and properties at the meso and micro scales. Performing VASP ab-initio calculations on automatically chosen sets of small models, MedeA-UNCLE captures the configurational complexity of real materials at different temperatures by means of Monte Carlo random sampling. Tight integration with job control in MedeA guarantees stability and fault tolerance. Graphical tools monitor progress of fully automated simulations and allow a ready-for-use visualization of results.
Key Benefits of MedeA-UNCLE:
• Models systems containing millions of atoms with DFT accuracy
• User-friendly setup within MedeA Environment
• Workflow-based automation of cluster expansion refinement
• Efficient handling of hundreds of input structures
• Intuitive graphical analysis and visualization
• Split and restart complex calculations
• Extend and expand existing Cluster Expansions
Computational Characteristics:
• Use Genetic Algorithm or Compressive Sensing
• Full integration with MedeA-VASP and other modules
• High throughput using the JobServer
Properties from MedeA-UNCLE:
• Structures of stable phases
• Vacancy concentrations
• Miscibility
• Random mixing energy
• Phase stability as a function of temperature and concentration
• Solubility
• Order-disorder transition temperature
• Micro structure
Target Audience: This one day workshop is intended for undergraduate project students, postgraduate students, postdoctoral researchers and researchers who are familiar with the field and want to employ state-of-the art methodology based on the density functional theory to understand bulk materials properties, phase diagrams, order- disorder transition temperature, vancacy concentrations, miscibility and micro structure.
Type of tutorial: Mix of tutorials and hands-on (mostly practical)
Special requirements:
• Attendees need access to a laptop or workstation, active CHPC user account cluster.
• Open source spreadsheet such as Microsoft Excel with graphing capabilities is required for post-processing of output data.
Outline of full syllabus:
1. Introduction to the MedeA® Software environment
a. Structure retrieval, building
b. Flowchart interface
c. High Throughput support
2. Introduction to Cluster Expansion techniques
a. Theoretical background
b. MedeA® Uncle overview
3. MedeA® UNCLE hands-on (bulk)
a. Selected tutorials
4. MedeA® UNCLE hands-on (surfaces)
a. Selected tutorials
5. Q&A, wrap-up
Description:
The goal of this workshop is to introduce students to CUDA and provide them with an understanding of parallel programming. CUDA is more than a few new keywords. One must understand SIMD and the pitfalls of serialization.
Students will leave with basic CUDA skills and some OpenACC knowledge plus useful machine learning and big data tools as well. My “from Hello World to exascale machine learning in one slide” will also be covered as data parallel training fits on GPUs nicely.
Target Audience:
Anyone with C/C++ programming skills in the Unix environment who wishes to learn about parallel programming and CUDA. The material will be 60% beginner, 30% intermediate, and 10% advanced.
Prerequisites:
C/C++ along with an ability to edit and compile programs in a Unix environment
Special requirements:
Users will have access to a CHPC system with GPUs.
Attendees should bring their own laptops. The ability to view pdf or PowerPoint files is required.
Outline of full syllabus:
08:00 Registration
09:00 Introduction and morning talk (30 minutes)
Login details and extracting the workshop material (15 minutes)
Section 01: Parallel intro and a first CUDA program)
Section 02: Profiling on a GPU
10:30 Morning Refreshment Break
11:00 Section 03: More CUDA and the Thrust Interface
Section 04: “From Hello World to TF/s machine learning”
12:30 Lunch
13:30 Afternoon talk
Section 05: Controlling parallel resources
Section 06: C++ objects and transparent host/GPU data movement
15:00 Afternoon Refreshment Break
15:30 Section 07: Task level parallelism on a GPU
Section 08: Managing big data, CUDA as a scripting language via dynamic load/link
17:00 End of Day
Additional Comments:
Students can work at their own pace.
Introductory students will learn the basics of CUDA and the profiler as well as how to think in parallel and understand the impact of parallel hardware on performance.
Intermediate/advanced students will hone their thinking about parallel programming and the limitations and advantages of GPU hardware. Extra credit exercises will challenge them.
All students will learn how to use machine learning and the ability to explore this hot field and leave with a tool that allows them to train and predict using their own data sets and neural network architectures. Further, they learn how to work with and collaborate using big data.
Where to begin? HPC as a field, a technology, and a tool for a diversity of disciplines that are enabled by it is an interplay of foundational concepts, rapidly evolving knowledge, and skill sets including programming. This one-day tutorial is presented as a beginners’ guide to supercomputing to expose the novice to the breadth of issues needed to begin to understand and use such systems for real-world problems. Included are basic concepts, HPC architecture, benchmarking, parallel programming using OpenMP and MPI, basics of GPU accelerators, and libraries. Participants will be shown live demonstrations of all of these techniques and given opportunities for hands-on experience using the CHPC Petaflops scale supercomputer in South Africa. Questions will be welcome from the attendees throughout the presentations.
Description:
This session will cover various aspects of analysing and optimising parallel programs to achieve optimum performance on the supercomputer. Tuning and Analysis Utilities (TAU) callgraph visualization system will demonstrate how to analyse different objects (e.g. modules, routines and functions) and identify performance bottlenecks within parallel applications.
Target Audience:
Parallel program users, developers and administrators
Prerequisites:
Attendees must at least have knowledge of compiling and running parallel applications on a cluster.
Type of Tutorial:
Mix of lectures and practicals
Special Requirements:
No need for attendees to have laptops. Presenter will demonstrate practical exercises on his laptop.
Outline:
08:00
Registration
09:00
10:30
Morning Refreshment Break
11:00
12:30
Lunch
13:30
Tuning and optimisation of parallel programs on the CHPC supercomputer
15:00
Afternoon Refreshment Break
15:30
Tuning and optimisation of parallel programs on the CHPC supercomputer
17:00
End of Day
CHPC Cyber Security Symposium (CCSS) – 1 full day
Topic: Understanding Risk in Shared CyberEcosystems workshop
As Technology continues to evolve so are the opportunities and challenges it provides. As a Society that relies largely on technology to bring us benefits, it also exposes its users to threats by the very nature of the opportunities it presents, thus becoming a focal point for Cybercrime, industrial espionage, and Cyberattacks. Protecting against these threats is of paramount priority.
Cyber Security domain is inherently dynamic, not only does system configuration changes frequently with new releases and patches, but also new attacks and vulnerabilities are regularly discovered. The core threat in Cyber Security is human, hence intelligent in nature. The attacker adapts to the situation, target environment, and countermeasures.
Attacks actions are driven by attacker’s exploratory nature, thought process, motivation, strategy, and preferences.
The goal of this workshop is to address some of the challenges faced by Network and Security Administrators in their institutions.
The workshop will address challenges such as:
• Iot Threats
• Blockchain Revolution
• Ransomware Evolution
• Future of Cybersecurity
• New hacking Methodologies
• Phishing
• Distributed Denial of Service (DDoS)
• Malware
• Internal Privilege Misuse
• Threats actors
• How to protect your environment
The following are our invited speakers:
• Prof Elmarie Bierman: Directory of SA Cyber Security Institute
• Dr Noelle Cowling : Stellenbosch University
• Dr Jabu Mtswene : CSIR
• Dr Noluxolo Gcaza: CSIR
• Mr Justine Westcott
• Mr Sikhumbuzo Mthombeni: Demension Data
• Mr Keti Cedric : Demension Data
• Mr Richard Hlalele: University of Johannesburg, Senior Manager ICT Strategy and Governance
The following are workshop Coordinators:
Mr Bigani Sehurutshi: University of Botswana
Mrs Lee-Anne Benjamin: Manager ,Corporate IT Compliance NRF
Address by Department of Science and Technology Director General, Dr Phil Mjwara
Address by Director of the CHPC, Dr Happy Sithole
In this keynote presentation, Trish will discuss the changing landscape of high performance computing, key trends, and the convergence of HPC-AI-HPDA that is transforming our industry and will fuel HPC to fulfill its potential as a scientific tool for insight and innovation. Trish will highlight not only key forces driving this shift but discuss how this transformation requires a fundamental paradigm shift and is opening up unprecedented opportunities for HPC.
Bioinformatics occupies the space between biology and computing and aims to answer
questions in biology using analytical and computing methodology. At the South African
National Bioinformatics Institute (SANBI), our research focus is on methods to store, retrieve
and analyze genetic information that spans both communicable and non-communicable
diseases. In the context of Public Health, that is a need to interrogate genetic information
(DNA) from both hosts (human) and pathogens (bacterial or viruses) to understand
susceptibility to diseases and ultimately to track infection trends in real time.
The ever-increasing volume of data being generated in the public domain places a strain on
in-house computing resources. While the computing facility at SANBI-UWC is adequate for
initial R&D, these resources are inadequate to complete projects timeously. The nature of
the bioinformatics workflows that require CHPC resources can be grouped into (1) high
throughput computing resources that are needed to describe the 1000s of genetic messages
in a genome, versus (2) high performance computing resources that is needed to model a
drug-protein interaction environment - these simulations require days/weeks of dedicated
compute time.
We have leveraged the CHPC facility in the context of infectious disease research with a
view to analyze genetic variation in bacterial genomes, and to identify drug targets in
pathogen genomes. These analyzes requires an environment that is able to support
reproducible workflows and virtualization of the software environment. Examples of these
use cases will be described.
Climate change is the most serious collective environmental challenge ever faced by modern humankind. It is a problem with global reach, but the research effort to address it is disproportionately concentrated in the northern hemisphere and in developed countries. Southern hemispheric and African climate issues differ from those that drive the research and modelling effort in the north. In particular, oceans dominate the southern hemisphere and the land is largely occupied by arid systems and tropical forests. African terrestrial ecosystems and processes, Southern Ocean physics, biochemistry and circulation dynamics as well as Southern Hemisphere atmospheric processes are under-studied and poorly represented in global models - despite being globally important contributors to earth system processes. In particular, of the about thirty currently existing coupled ocean-atmosphere global circulation models (CGCMs) and Earth System Models (ESMs) suitable for the projection of future climate change, only one model had its genesis in the Southern Hemisphere. Towards addressing this disproportionality, and in alignment with the South African Department of Science & Technology's Global Change Grand Challenge, the CSIR and partners are invested in building a Variable-resolution Earth System Model (VrESM), with the aim of contributing projections of future climate change to the Coupled Model Intercomparison Project Phase 6 (CMIP6) and Assessment Report 6 (AR6) of the IPCC. VrESM is the first African-based Earth System Model (ESM) and has as component models the variable-cubic atmospheric model (VCAM) of the CSIRO, a dynamic land-surface model (CABLE), the variable cubic ocean model (VCOM) and an ocean biogeochemistry model (PISCES).
VrESM is formulated on the non-orthogonal, quasi-uniform cubic grid of Purser and Rancic (1998). Its atmospheric component VCAM has been developed by the Commonwealth Scientific and Industrial Research Organisation (CSIRO). VCAM is a σ-coordinate model that uses a semi-implicit semi-Lagrangian approach to solve the hydrostatic primitive equations. VCAM has inherited the comprehensive physical parameterisations of the Conformal-Cubic Atmospheric Model (CCAM) (McGregor, 2005b). The dynamic land-surface model used is the CSIRO Atmosphere Biosphere Land Exchange model (CABLE). CABLE includes a dynamic river routing scheme adapted from the CSIRO Mk3.5 CGCM. The VrESM ocean component VCOM has been developed by the CSIR. This model solves the Boussinesq hydrostatic equations in a z-coordinate in momentum-conservation form, using a split-explicit solution procedure. VCOM is coupled to the PISCES ocean biochemistry model. Coupling of the ocean, atmospheric and land-surface components takes place every time-step. It is envisaged that VrESM will be applied on a 100 km horizontal resolution grid within CMIP6, with a longer-term plan of performing global eddy-resolving (10 km resolution) simulations depending on the availability of supercomputing resources at the Centre for High Performance Computing (CHPC) of the CSIR. Our path over the next 5 years (towards and beyond AR6) is firstly technical, in further developing and optimising the VrESM on the computer cluster at the CHPC and secondly scientific, in understanding the underlying mechanisms that are often parameterized rather than resolved in global climate models – followed by subsequent improvements and optimizing the model with this new knowledge. In particular, South Africa has invested significantly in resources to investigate and undertake long term observations that resolve critical dynamics in the oceans to our south. From this investment has emerged a competitive edge in building the first ESM that realistically represent Southern Ocean biochemistry and the role of the Southern Ocean in the African and global climate systems. In this presentation, we will present the unique aspects of the VrESM numerical solution procedure as applied on the cube-based grid. A variety of model applications (including simulations of present-day and future climate) will also be shown, with a focus on Southern Hemisphere and African climate processes.
Physical techniques for surface modification of plastics use surface-active agents, which are allowed to self-assemble at the surface. Many techniques, which are important in modern technologies, use polymer blends, and there is considerable interest to understand the extent the composition of the surface layer differs from that in the bulk for molten polymer mixtures. Dynamical and structural properties of polymers in the melt state are strongly influenced by molecular architecture [1-4] and blending polymers with different molecular topologies could be potentially exploited to control interfacial segregation of the polymer film, and to achieve optimal mechanical properties of the plastic material [5,6]. However, a deep understanding of the role of chain architecture and molecular mass in determining which species preferentially adsorb at a given interface is lacking. Experiments to resolve the matter are typically conducted by mixing polymers possessing the same repeat chemistry, but different molecular architecture [10–14]. Here we show the results obtained in large-scale molecular dynamics simulations of linear-cyclic polymer films, and we find clear evidence of enhancement of linear polymers at the interface [7], in agreement with recent experimental results [8]. The behavior predicted by the self-consistent field theory (SCF), i.e., enhancement of cyclic polymers at the interface [9], emerges for relatively long chains. In our presentation, we provide a picture of the microscopic mechanisms through which the chain length arbitrates the competition between the different packing constraints imposed by the loop and linear geometry of the two polymers. We also discuss the role of enthalpic and entropic factors of the interfacial free energy of the system in determining which species in the blend preferentially adsorbs at the interface.
[1] Kapnistos, M; Lang. M; Rubinstein, M; Roovers, J.; Chang, T; Vlassopoulos, D., Soc. Rheol. Annu. Meeting 2006.
[2] Robertson, R. M.; Smith, D. E., Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 4824-4827.
[3] Iyer, B. V. S.; Lele, A. K.; Shanbhag, S., Macromolecules 2007, 40, 5995.
[4] Subramanian, G.; Shanbhag S., Macromolecules 2008, 41, 7239-7242.
[5] Wu, S. L.; Polymer Interface and Adhesion, Marcel Dekker: New York, 1982.
[6] Garbassi, F; Morra, M; Occhiello, E., Polymer Surfaces: From Physics to Technology; John Wiley and Sons: New York,1994.
[7] G. Pellicane, M. Megnidio-Tchoukouegno, G. T. Mola, and M. Tsige, Physical Review E Rapid Communications, 93, 050501 (2016); M. Megnidio-Tchoukouegno, F. M. Gaitho, G. T. Mola, M. Tsige, and G. Pellicane, Fluid Phase Equilibria, 441, 33–42 (2017).
[8] Wang, S-F; Li, X.; Agapov, R. L.; Wesdemiotis, C.; Foster, M. D., ACS Macro Letters, 2012, 1, 1024-1027. (2016).
[9] Wu , D. T.; Fredrickson, G. H., Macromolecules 1996, 29, 7919-7930.
Due to increasingly global environmental and energy crises, visible light semiconductor photocatalyst with a tunable bandgap and optical properties have received attention. For this reason, developing efficient and cost–effective photocatalysts for environmental remediation is a growing need, and semiconductor photocatalysts have now received more interest due to their excellent photocatalytic stability and activity. The charge transfer, catalytic stability, electronic and optical properties of several semiconductor–based photocatalyst materials were systematically studied using first–principles study. All the calculations were performed using the Cambridge Serial Total Energy Package (CASTEP) code [1] implemented in Materials Studio 2016 [2] with the plane–wave ultrasoft pseudopotentials method [3] and Perdew–Burke–Ernzerhof (PBE) functional for the exchange and correlation contribution [4]. All the simulations were done using the resources provided by the Centre for High Performance Computing (CHPC), Rosebank, Cape Town [5]. The proposed photocatalyst materials show high photocatalytic activity under visible light irradiation with good stability and reduced bandgap compared to the bulk semiconductor. The heterostructures formed a type−II band alignment to accelerate the interfacial charge transfer process and the photocatalytic activity. By comparing the relative ratio of effective mass and band alignment results, we could conclude that heterostructures have not only superior mobility of charge carriers, but also higher separation of photoinduced electrons and holes.
References
[1] M. Segall, P.J. Lindan, M.a. Probert, C. Pickard, P. Hasnip, S. Clark and M. Payne. First–principles simulation: ideas, illustrations and the CASTEP code. Journal of Physics: Condensed Matter 14 (2002) 2717–2744.
[2] Materials Studio simulation environment. Release 2016, Accelrys Software Inc, San Diego, CA (2016)
[3] D. Vanderbilt. Soft self–consistent pseudopotentials in a generalized eigenvalue formalism. Physical Review B 41 (1990) 7892–7895.
[4] J.P. Perdew, K. Burke and M. Ernzerhof. Generalized Gradient Approximation Made Simple. Physical Review Letters 77 (1996) 3865–3868.
[5] Center for High Performance Computing. CHPC.
The use of quantum sensors to investigate gravity, dark matter, and the early universe is in the vanguard of a 2nd Quantum revolution; as significant as the first deployment of telescopes it will transform the way we understand the world. The technological innovation that is the engine of society’s development has been initiated and fuelled by fundamental scientific research; from Faraday’s work on electricity to the development of the world wide web. In the 20th century the application of our best understanding of the sub atomic world – quantum mechanics- generated new knowledge about the world and new technologies that improve the human condition. Examples include semiconductor microelectronics, photonics, the global positioning system (GPS), and magnetic resonance imaging (MRI). These technologies underpin significant parts of the economies of developed nations, we refer to this as the “1st Quantum Revolution”. Future scientific and technological discoveries from the application of quantum mechanics may be even more impactful – a 2nd Quantum Revolution. The areas that will be potentially transformed include biology, the defence sector and fundamental science. It is the latter that is the focus of this talk.
Quantum machine learning investigates how quantum computers can be used for data-driven prediction and decision making. The talk introduces to this relatively young discipline and shows the potential of "Big Data" applications on near-term quantum computers, as they can be found in the cloud at present. Data encoding into quantum states, quantum algorithms and routines for inference and optimisation, as well as the construction and analysis of genuine ``quantum learning models'' will be introduced.
Still in early development, quantum computing is already overturning our contemporary notions of computational methods and devices. Using new concepts of computing based in quantum physics, quantum computers will be able to solve certain problems that are completely intractable on any imaginable classical computer, such as accurate simulations of molecules and materials, or breaking public key encryption. I will describe the hardware and software architecture of quantum computers and Microsoft’s full stack approach, from novel topological quantum bits to real-world quantum applications. I will attempt to dispel myths and hype surrounding the field and present a realistic assessment of the potential of these devices, and the specific application areas on which they are expected to have a large impact.
To be confirmed
The convergence of AI and HPC has created a fertile venue that is ripe for imaginative researchers - versed in AI technology - to make a big impact in a variety of scientific fields. From new hardware to new computational approaches, the true impact of deep- and machine learning on HPC is, in a word, “everywhere”.
Just as technology changes in the personal computer market brought about a revolution in the design and implementation of the systems and algorithms used in high performance computing (HPC), so are recent technology changes in machine learning bringing about an AI revolution in the HPC community. Expect new HPC analytic techniques including the use of GANs (Generative Adversarial Networks) in physics based modeling and simulation as well as reduced precision math libraries such as NLAFET and HiCMA to revolutionize many fields of research. Other benefits of the convergence of AI and HPC include the physical instantiation of data flow architectures in FPGAs and ASICs plus the development of powerful data analytic services.
The research is based on the development and design of metal alloys with special applications in various industries, from automotive and medical to aeronautical. These alloys have unique properties and behaviour under pressure and temperature conditions. Focus will be on the precious metal, iron-aluminium, titanium and titanium-based systems for shape memory alloys, as well as zirconium-niobium systems. A density functional theory based semi-empirical approach is employed to explore material properties at zero temperature; while molecular dynamics techniques are also carried out to predict the stability, strength, behaviour and the extent of transformation temperature. The overview on computational codes and usage of HPC will be discussed. Furthermore, the advances on machine learning approaches towards titanium cluster development and growth will also be indicated.
The burning of fossil fuels is identified as the biggest source of Sulfur dioxide (S02) emissions in South Africa. SO2 is a colourless gas with a nasty smell. It impacts both human health in the form of diminishing lung function and the environment through the effects of acid rain leading to deforestation. In order to meet minimum emissions standards for S02, as provided by the Air Quality Act of South Africa, big emitters such as Eskom and Sasol are required to implement Flue Gas Desulfurization (FGD) plants. Current FGD plants in South Africa are wet operated – that is, a sorbent like Calcium Hydroxide ( Ca(OH)2 ) is mixed with water and sprayed as slurry into flue gas. In order to save water, dry FGD processes are considered. One such process is called Dry FGD using a Circulating Fluidized Bed (CFB). During this process dry Ca(OH)2 is introduced into a flue gas stream while water is separately introduced as a fine spray. SO2 is then reduced in a reaction involving the flue gas, Ca(OH)2 particles and water droplets.
The main reactor of a Dry FGD CFB plant is the vertical riser, which can typically be 3m in diameter and 20m high. The riser acts as a container in which the SO2 in the flue gas is exposed to the Ca(OH)2 particles and water droplets. Of importance are the residence time and the riser hold-up, which forms part of the overall successful design of a dry SO2 scrubber. Computational Fluid Dynamics is used to model the gas-solid two-phase mixture. The two-phase flow regime extends from dense to dilute turbulent flow, which renders the solution especially challenging. A Probability Density Function (PDF) approach is used to represent the microscale dynamics while filtering methods are used to yield practical solution on the scale of the plant.
Two codes are used for modelling the two-phase flows namely Neptune_CFD, which is the research code partly owned by EDF in France, and OpenFOAM, which allows the user to make fundamental changes in the source code. Both these codes are run at CHPC. The results of the study presented focus on transient runs performed on CHPC. Optimization of nodes and cores were done to yield the most effective solver parallelization for these codes. Hold-up and resident times of particles are predicted inside the riser. The mathematical models of the riser application of two-phase flows represent from the biggest computerized parallel models in this field.
Bit patterned media is one of the promising approaches to extend magnetic recording densities in hard-disk drives beyond recording densities of one Terabit (10^{12}) per square inch. In this approach, the magnetic medium is patterned into nanometer-sized magnetic islands where each island can be considered to represent a binary digit. The islands are magnetically isolated but experience magnetostatic interactions between them. This study investigates magnetostatic interactions between islands of various shapes for various island separations using micromagnetic simulations. The shapes range from truncated elliptic cones to cylinders. The computation of magnetostatic interactions is a major time-consuming task in a micromagnetic simulation. These interactions scale with O(N^{2}) operations, where N is the number of interacting mesh elements in discretized islands. To carry out simulations, an open-source finite element micromagnetics package called magpar was used. The island mesh was generated using netgen, a tetrahedral mesh generator. An open source visualization tool called paraview was used view the outputs of simulations. Open MPI was used in the simulations. Twelve cores on one node were used in this work. The interactions between islands have been compared against dipole-dipole interactions where each island is assumed to be a dipole. The study has shown that for islands considered, the effect of island shape is important for island separations less than twice the island width, centre to centre. The dipole approximation is only sufficient for island separations beyond twice the island width. This result further suggests an improvement in the computation of time-consuming magnetostatic interactions between large numbers of islands by treating distant islands as dipoles.
Container technology offers a convenient way to package an application and supporting libraries such that moving them from platform to platform can be done without having to rebuild. Additional features, such as stateless execution enable restarting a containerized application with minimal penalty elsewhere. Combining better support for storage into the container ecosystem breaks this stateless model, but offers advantages. This talk will examine the advantages and penalties of this approach and offer solutions to ease adopting the idea.
Many important HPC applications are communication-bound and/or I/O-bound. These applications depend on efficient inter-process communication and I/O operations, hence, network interference can cause significant performance degradation. Unfortunately, most modern HPC systems use the same network infrastructure for both MPI and I/O traffic, with multiple jobs sharing the system concurrently. The scarcity of studies that investigate the interference between MPI and I/O jobs leaves us with only a vague understanding of how these types of traffic interact with each other; the interference characteristics are not well explored and neither are the strategies for avoiding this interference in order to improve performance.
In this talk, we discuss the important characteristics of the interference between I/O and MPI traffic on fat-tree networks, exposing the impact of factors such as message size, job size, and communication frequency on the resulting interference. We show the extent to which MPI traffic is more sensitive to interference than I/O traffic on a fully provisioned fat-tree network, and we categorize configurations that can cause even an I/O job to be slowed by 1.9X due to interference from MPI traffic. This work has pinpointed the most significant aspect of the performance trends: the I/O-congestion threshold. This threshold refers to the frequency of sending I/O requests when MPI jobs start experiencing detrimental performance degradation due to I/O interference while, simultaneously, I/O traffic becomes relatively insensitive to MPI interference.
The insights gained from the interference characterization can be used with knowledge of the network topology to mitigate the effects of this inter-job interference on application performance. Our work shows how careful placement of jobs and I/O servers can, independently, mitigate interference. Additionally, I/O throttling can be guided by the I/O-congestion threshold to improve MPI performance by up to 200% while incurring only a 18% slowdown in the I/O performance.
Through the first several decades of computing, two data storage abstractions/paradigms dominated common practice: Files and relational databases. While there is significant potential overlap between their use, it is often easy to decide which is more efficient for a particular application or workload. However, over the last twenty years, teh rise of new patterns for parallel and distributed computing (i.e., those now loosely grouped under the "cloud" umbrella) have brought to the fore other kinds of storage technologies and techniques. Among this new group of options are NoSQL stores, object stores, and key-value stores.
The lines between these storage technologies are blurred, both because of overlapping application benefits and different design choices for various types of packages. What kinds of workloads benefit most from each paradigm? Who might an HPC user leverage these technologies? This talk will address these questions and more.
Fresh from the SC18 conference, Addison Snell of Intersect360 Research will give a market overview presentation of developments and predictions for HPC, Hyperscale, and AI. Views will include the latest market forecasts, and a rundown of the big news from SC18 — what was hot, and what was not.
Following the presentation, Addison will transition directly into the fast-paced CHPC Vendor Showdown panel, an audience favorite, in which representatives of participating vendor companies will answer pointed questions, head-to-head, and be awarded points based on who answered the question the best for the audience.
The presentation will focus on the key aspects of the report of the World Bank on "The Role and Status of National Research and Education Networks in Africa" [1], which are applicable universally. The aim of this report, which is an open education resource, was to provide guidance to governments, institutions, and development partners on how to approach the provision of advanced information and communication technology (ICT) services to the higher education and research community in Africa. The premise of the report that the organization of ICT services and connectivity is best provided by a dedicated organization called the National Research and Education Network (NREN) is based on international best practice and is applicable worldwide.
Beginning with a brief overview of progress to date in all regions in Africa, including North Africa, the presentation will dwell more extensively on the important but often invisible role that NRENs play in the creation and the sharing of knowledge. NRENs not only provide the essential connectivity services to higher education institutions, but they do it uniquely within a global community that has its own rules and agreements for access and identity. In this way they provide customized services to the academic community that are not part of the offering of commercial Internet Service Providers (ISPs). The creation of an NREN in a country, that is also connected to its neighbours and globally, is a vital step to break through the academic isolation that universites and colleges may have experienced in the past and it can be the catalyst to be invited to join international research consortia. Creating new knowledge through research is now a thoroughly collaborative global endeavour, but partipation is limited to those with adequate connectivity. And teaching and learning itself, the ‘E’ in NREN, has moved beyond dependence on textbooks and local lectures to an exercise in independent learning based on researching multiple sources of knowledge, which are primarily in digital form and accessible through electronic networks.
The presentation will outline the case for the establishment of an NREN, along with the services that it may provide, with possible governance models, and the various ways it can be made financialy sustainable. Finally, the trends in the use of ICT in higher education and how they might affect the operation of NRENs are discussed as a prelude to guidance on how to go about establishing or strengthening an NREN, with recommendations to the government, the private sector, institutions, and development partners.
[1] https://openknowledge.worldbank.org/handle/10986/26258
The advent of the fourth industrial revolution results in devices, endpoints and sensors continually being added to extract and gather data from networks, systems and people. Ensuring security on all these entities is vital, but more so the ability to move from data to information to knowledge and finally intelligence in this big data environment. In terms of the cyber warfare space, power will be in the hands of those who control information and that can extract, analyse and apply security intelligence in a dynamic environment.
In his 2002 State of the Nation Address, President Thabo Mbeki identified Information and Communication Technology (ICT) as “a critical and pervasive element in economic development” and recommended the establishment of an “ICT University”. The establishment of the CHPC in 2005 signalled the first pillar in the national research cyberinfrastructure1 and remains a major investment that anchors the development and growth of the computational and data sciences stretching across disciplines from science, through to social science. The recent feverish announcements from pundits through to politicians hailing the onset of the 4th industrial revolution posits a research landscape where computational science is a central kernel of South Africa’s future economic plan.
Large parts of the chemical, biological and physical science have had an uptick in production and activity in modelling, simulation and data analytics as measured by number of papers produced. However, the challenge of innovative home grown methods to address the national needs of medical science, the bio-economy and the green revolution remain unmet.
I will present a view of the future of computing and data analytics in the biological, chemical and life sciences. More importantly I will discuss a hopeful future scenario where academics across all South African institutions cooperate via the CHPC to advance our international research profile.
During the past decade, massive parallel DNA sequencing technologies have completely changed the way in which genetic data are generated and analysed. Instead of sequencing a few hundred nucleotides and focusing on a handful of genes, it is now possible to generate data from entire genomes. The human genome contains about 3 billion nucleotides, and even this is tiny compared to some plant genomes. The resulting datasets are so large that it is impossible to assemble and analyse them without HPC. In this presentation, I demonstrate some of the HPC pipelines my lab uses to assemble genomic data and reconstruct evolutionary relationships between species. These data sets are small compared to what we have planned for the near future, highlighting the necessity to invest more heavily in HPC resources in South Africa and help the country’s biologists become internationally competitive. I also focus on the present reluctance of many traditional geneticists to adopt the new technology, and suggest that this bottleneck needs to be closed by finding common ground between biologists, computer scientists and other stakeholders. In the near future, we need to jointly train bioinformatics students who operate at the interface of the different disciplines.
Pyrometallurgy refers to high temperature (>1600°C) extraction of valuable metals from mineral ore deposits. Most pyrometallurgical processes occur in furnaces of various types in which many complex phenomena are occurring simultaneously. Almost none of these phenomena can be observed directly due to the extreme conditions inside furnace vessels. Computational modelling is invaluable in understanding the processes in and around furnaces and is often used for design and optimization. Problems vary from simple CFD analyses to coupled multiphysics applications such as magnetohydrodynamic models of plasma arcs.
High performance computing is a valuable tool for developing increased understanding and better engineering for these challenging problems. A few case studies are given: DC furnace arc modelling, multiphase flow through furnace tap-holes, combustion in a rotary converter and settling of metal and slag.
The Paris Agreement which was achieved in December 2015; holds signatory countries responsible for keeping the increase in global average temperatures well below 2°C with respect to the to the pre-industrial period and to strive to limit the temperature increase to 1.5°C , recognising that this will reduce the impacts of climate change. Given this impetus at global level, it is of paramount importance to consider the implications of the 1.5 , 2, and 3 °C thresholds in the tropical cyclone activities within the South West Indian Ocean Basin.Using the Coordinated Regional Downscaling Experiment-Africa (CORDEX) regional climate models, we downscale six global climate models of the Coupled Model Inter-comparison Project Phase 5 (CMIP5) to high resolution with the aid of computing power from the south African (CHPC) Centre for High Performance Computing’s Lengau Cluster.This serves towards studying changes in tropical cyclone tracks over the South West Indian Ocean under different extents of global warming(1.5, 2 and 3° C of warming with respect to pre-industrial conditions). It is projected that the number of tropical cyclones making landfalls over southern Africa under global warming will decrease, with 2 °C being a critical threshold, after which the rate of cyclone frequency with further temperature increases no longer has a diminishing effect.
Next Generation Sequencing has brought genomic analysis within the range of a great number of laboratories, while increasing the demand for bioinformatic analysis. These typically comprise workflows composed out of chains of analyses with data flowing between workflow steps. Such analysis is amenable to High Throughput Computing, a form of high performance computing characterised by a focus on overall analysis throughput rather than optimisation of a single application. In recent years workflow languages and container technologies have become a key part in composing efficient, reproducible and re-usable bionformatic workflows. These technologies, however, pose a challenge for High Performance Computing providers as they require different characteristics from an execution environment to that provided by traditional HPC clusters. These challenges will be discussed and some approaches to solving them will be discussed.
As part of hosting the Square Kilometer Array (SKA) mid frequency radio telescope in the Northern Cape Karoo region, the South Africa Radio Astronomy Observatory (SARAO) will be providing suitable facilities to house the computing, networking and data storage for both the Science Data Processor (SDP) and an SKA Regional Science Centre (SRC). These two facilities are expected to host petascale computing, exascale storage capacity and network connectivity in order of 8 Tbps by 2024. The facilities are expected to be located in Cape Town, over 700 kms from the Central Signal Processor (CSP) at the Telescope core site near the town of Carnarvon.
The purpose of this presentation is to share the latest information that will be used to define requirements for the construction of the two co-located facilities, each with their own governance models, operation models and potentially unique approaches to fulfilling the computing infrastructure requirements.
Currently SARAO is working together with the SKA Organisation (SKAO), the Department of Science and Technology (DST) South Africa and the Western Cape Government (WCG) on initial plans and aims to start to journey of engaging with other stakeholders, gathering information and improving accuracy of detailed information with the aim to assist with the planning that will ultimately lead to construction of the required facilities.
Moving masses of data is a challenge. In most cases networks optimized for business operations are neither designed for nor capable of supporting the data movement requirements of data intensive research. When scientists attempt to run data intensive applications over these so called “general purpose”/enterprise networks, the result is often poor performance – in many cases poor enough that the science mission is significantly impacted. At its worst this means either not
getting the data, getting it too late or resorting to “desperate” measures such as shipping disks around. The South African National Research Network (SANReN) is currently piloting a data transfer service with the goal of changing this for our researchers/scientists and optimising the transfer of datasets across the network. The service makes use of data transfer nodes configured in a science
DMZ architecture using specially designed data transfer tools to efficiently and securely move data.
This presentation will present an overview of the science DMZ, data transfer nodes, tools and services that we have implemented as part of the SANReN Data Transfer Pilot project as well as preliminary results we’ve achieved.
In South Africa, increased access to educational resources is altering the collective face of education. Many in the education field see new technologies as powerful tools to help universities meet the needs of ever-more-diverse student populations. Digital devices, software applications, interactive digital textbooks, e-books, online assessments and various learning platforms offer unimaginable options for tailoring education to each individual student’s academic strengths and weaknesses, personal preferences and financial position. Without a doubt, one key enabler for these educational technologies is fast, robust and reliable broadband. Reliable broadband not only assists on-site students and lecturers at university campuses, but also enables off-site faculty and students to participate in online learning experiences. Our increased dependence on these online learning technologies will continue to increase the demand for more and more bandwidth.
This talk looks at the staggering numbers surrounding the use of fast, robust and reliable broadband at tertiary educational institutions. We consider the growth in Internet traffic at educational institutions, bandwidth costs and the advancement of educational technological tools and media
Social Engineering has become the go-to attack methodology in the 21 st Century. Organisations and governments have a vested interest in securing sensitive information as well as the trust of clients or citizens. Technology on its own is no longer a sufficient safeguard against information theft. Employees, often the weak link in an information security system, could be influenced or manipulated by the attacker to divulge sensitive information, thus allowing unauthorised individuals to gain access to protected systems. As technology developed, social engineering became commonly used for technology related crimes. This is no longer the case as we see it reverting back to its original use of mobilising communities to perform or act in ways assumed to be ‘optimal’ or ‘correct’. This is typically either for personal gain of the attacker or to influence people to subscribe to the attacker’s societal belief system.
Through the utilisation of real-life social engineering examples, the formal process of social engineering and techniques utilised to abuse the inherent trust of individuals is explored. This particular discussion focuses on the psyche of an individual and discusses the basic human instinct of trust. The concept of trust is something that differs based on environmental conditions and socio-economic upbringing. Trust will however, always remain a basic human instinct. A ‘functioning’ society relies on the fact that individuals have inherent trust. For example, most people inherently ‘trust’ institutions to manage financial aspects of life, similarly, they “trust” that other people in turn will obey those same institutions. Not having this basic form of trust could send society into a state of chaos. As a result, this discussion examines the extent to which
scepticism can aid society and in turn allow individuals to protect themselves. In the end, everyone has a need of being accepted and ultimately, all of us just want to belong…
The main research interest of the Research Unit in Bioinformatics (RUBi) at Rhodes University is in structural bioinformatics and its applications to bio-economically important research questions, i.e. to drug discovery projects for diseases related to Africa, and to biodiversity and bioprocessing. Centre for High Performance Computing (CHPC), South Africa, is regularly used for accelerating computer intensive methods including drug virtual screening, molecular modelling, quantum mechanics, molecular mechanics, molecular dynamics (MD) and combinations of these techniques. RUBi’s research is computationally highly expensive; i.e. within 3 months (February – April 2018) a total of 7.2 million cpu hours was used. Hence, CHPC has a great impact on RUBi’s research, international collaborations and human capacity development. This talk focuses on the application of HPC to RUBi’s research projects with some specific examples including analysis of drug resistance of HIV, identification of new drug targeting sites for cancer, and investigation of proteins related to the production of biofuels. Further, computational details, e.g. benchmarking, compute time, details of software used and/or developed will also be given to demonstrate the need for HPC resources.
The unique nature of the underwater mining environment has resulted in R&D playing an import role within De Beers Marine. Physical and numerical simulation methods form an integral part of the R&D concept development pipeline. Computational Fluid Dynamics has been increasingly used for concept exploration, risk mitigation as well as to enhance understanding of physical simulations. HPC has been an enabler for using Computational Fluid Dynamics to solve multi-phase problems that were previously deemed to be too large and/or complex. Most recently, this capability has been used in the feasibility stage of a new marine mining vessel.
Weather forecasting using Numerical Weather Prediction (NWP) models is a well-established science and based on the fundamental equations of fluid dynamics. NWP is therefore an important tool for predicting weather and making climate projections. The grid spacing used in models is largely dependent on the available computational resources and processing speed required for model output availability for timeous decision making. The direct numerical simulations that are able to capture all processes explicitly are still too far from our reach for practical purposes such as NWP, seasonal forecasting and climate change studies.
The available computational resources have been improving over time, which have made possible a decrease in the grid spacing used in models. A number of meteorological organisation are now using grid spacing of 4 km or less for NWP purposes, especially over limited areas. Models used with such high resolution are called Cloud Resolving Models (CRMs) or convective scale models, and in these models clouds are thought to be resolved explicitly. Similar to other meteorological organisations SAWS runs their NWP models with a grid spacing of 4.4 km over southern Africa and 1.5 km over South Africa. Both these configurations are made four times a day, but with different lead times. These simulations are made on a CRAY XC30 machine which was procured in 2014.
The SAWS CRAY XC30 has 168 nodes with Ivybridge processors. Each model simulation (4 km and 1.5 km) utilizes 1728 cores to complete. The CRAY is also used for operational seasonal forecasting with a coupled atmosphere-ocean model and there are plans for more models on air quality, ocean wave forecasting and applications research that will be undertaken on the CRAY. With all the operational applications on the CRAY, there is limited processing time for research activities such as sub-kilometre grid spacing simulations or ensemble forecasting. Big global operational centres such as the European Centre for Medium Range Weather Forecasts (ECMWF) maintain two HPC resources, with the second one available as a backup to their operational system. Due to budget constraints and the high costs of HPC resources, SAWS does not have such a failover system in-house, resulting in no or limited support for operational NWP at SAWS to operational forecasters if the operational system fails.
In order to deal with HPC shortcomings at SAWS a cooperative partnership agreement was entered into with the Centre for High Performance Computing (CHPC) to facilitate mutual HPC activities, and training programs to meet the missions of both SAWS and the CHPC. The CHPC is one of the three national cyber-infrastructure pillars that are supported by the South African Department of Science and Technology (DST). The CHPC currently hosts a Dell cluster, with a total of 1358 nodes with Intel v3 Haswell processes, and a 4PB Lustre storage. The main objective of the CHPC is to enable South Africa to become globally competitive and to accelerate Africa’s socio-economic upliftment through the effective application of high-end Cyberinfrastructure. The CHPC seeks to become an accomplished and preferred partner for High Performance Computing solutions.
The agreed areas of cooperation between SAWS and the CHPC include 1) the use of the CHPC cluster as a fail-over system for SAWS operations, 2) the use of the CHPC HPC system for benchmarking purposes to determine future operational needs of SAWS, 3) Use of the CHPC HPC system for research purposes, and 4) Training on the use of the CHPC cluster. SAWS has been using UK Met Office Unified Model (UM) as its main Numerical Weather Prediction model since 2006. As a result of the agreement between SAWS and the CHPC, the UM has now been installed on the CHPC cluster. The CHPC cluster was used successfully for business continuity purposes when SAWS moved offices, as well as, as a failover system when there were power issues at the SAWS premises. Before now the UM had only been available at SAWS and to SAWS employees due to license restrictions, resulting in scientists only getting exposure to the UM once employed by SAWS. The installation of the UM on the CHPC infrastructure will allow SAWS and other academic and research institutions across South Africa access to the UM for research purposes.
This contribution is aimed to illustrate how chemical computational results, obtained by density functional theory calculations on high performance computers, provides insight into many observed chemical phenomena, mysteries or questions, such as the following:
The structure of water clusters in bulk water is considered one of the unsolved problems in chemistry. How does the structure of water clusters look like at nanoscopic scale? How do the connecting hydrogen bonds render water ideally suited to life processes, being easily formed but not too difficult to break? A photochromic mercury compound has an orange colour in solution, but turns blue when sunlight shines on it. Why? How do these orange and blue forms differ in structure? Why is conductivity much higher for a certain iridium complex, than for the related rhodium complex with the same ligand? Why is the transient reaction intermediate within a multi-step red blood cell reaction not experimentally observable?
These and many other puzzling chemical problems, can be addressed very efficiently without experimental error, by computational chemistry conducted on high performance computers, thereby bypassing tedious or expensive laboratory techniques.
References:
[1] https://en.wikipedia.org/wiki/Water_cluster
[2] A. Malloum, J.J. Fifen, J. Conradie, The Journal of Physical Chemistry, submitted, MS no jp-2018-08976q.R1
[3] H. Schwoerer, K.G. von Eschwege, G. Bosman, P. Krok, J. Conradie, European Journal of Chemical Physics and Physical Chemistry, 2011, 14, 2653-2658.
[4] K.G. von Eschwege, J. Conradie, J.C. Swarts, Journal of Physical Chemistry A, 2008, 112, 2211-2218.
[5] J. Conradie, Journal of Organometallic Chemistry, 2017, 833, 88-94.
[6] S. Basu, R. Grubina, J. Huang, J. Conradie, Z. Huang, A. Jeffers, A. Jiang, X. He, I. Azarov, R. Seibert, A. Mehta, R. Patel, S.B. King, A. Ghosh, M.T. Gladwin, D.B. Kim-Shapiro, Nature Chemical Biology, 2007, 3, 785-794.
TensorFlow is the system driving Google's ML efforts. Many components make up this system, including a sophisticated user-friendly development environment, highly optimized language features and compilers, ultra-high performance custom chips called Tensor Processing Units (TPU), and scalable deployment on the world's devices. TPU pods may well eclipse traditional performance boundaries of the top HPC systems at a much lower cost. We will review TensorFlow's SW and HW environment, which begs the question of how usable it might be for HPC.
One of the prominent trends in computing is the convergence of supercomputers and embedded control computers, which have come to share many of the same requirements and limitations. These common attributes include multicore, power, reliability, programmability, and portability. The increasing use of lightweight processors like embedded cores in HPC systems prompts the need to unify multiple cores for time dependent embedded control. The challenges arising due to asynchrony of parallel execution, especially important in the context of non-homogenous many-task programs, make workload scheduling for optimal performance and predictability of overall execution time particularly difficult. This talk presents results of an NSF sponsored research project attempting to span the gap between the two classes of computer system through the conceptual bridge of a new execution model and its surrogate runtime system software and programming interface. It extended the ParalleX execution model to the domain of embedded computers by incorporating real-time semantics as an intrinsic property of the model so that multicore embedded computer architectures may be treated as a single system exploiting dynamic adaptive techniques to achieve real-time capability even when concurrent processing is required to reduce response time to necessary bounds. Performance figures derived using initial implementation of real-time extensions to the HPX-5 runtime system will also be presented and discussed. Questions from the audience are welcome throughout the talk.
Description:
The last session on Women in High Performance Computing in South Africa was proposed and held during 2017 annual conference. The major aim of the initiative was to establish a network of Women in HPC in SA, by bringing them together during the meeting. The session was supported and attended by both men and women, and most importantly was supported by CHPC management team. An interim steering committee comprising of women from various institutions and disciplines was established, and will take the initiative going forward. Consequently, we propose to have a session during 2018 annual conference.
It is aimed that at the end of the session and beyond, the following goals would be achieved:
• Being part of the solution: instructions for advocates and allies.
• Best practices from organizations on improving workplace diversity.
• Managing the two body problem and achieving effective work-life balance
• Encourage more female learners and students to consider careers in HPC
• Contribute in increasing the number of women and girls participation in HPC through training and networking.
• Share information and resources that foster growth for women in HPC.
Both men and Women are welcome to attend.
To be added
On the 13th July 2018 the Deputy President of the Republic of South Africa, David Mabuza, unveiled the MeerKAT telescope array.
The MeerKAT telescope is a mid-frequency 64-antenna array radio telescope. The MeerKAT telescope generates a large amount of data. The Science Data Processing subsystem is responsible for ingesting the telescope data, cleaning the data, packaging that data for scientific analysis, delivering that data to science groups, and storing that data for future use.
In order to achieve these objectives, the Science Data Processing team has had to innovate in how to build the appropriate high performance computer systems to meet MeerKAT telescope performance and capacity requirements. Project constraints included budget, human resources, technology availability, and problem novelty.
Khutso Ngoasheng will narrate the audience through the journey to delivering the MeerKAT telescope's Science Data Processor. He will talk briefly about current activities and challenges within the Science Data Processing subsystem. He will also introduce potential future paths for the Science Data Processor, especially in the Square Kilometre context.
Big Data analytics is the often complex process of examining large and varied data sets -- or Big Data -- to uncover information including hidden patterns, unknown correlations, market trends and customer preferences that can help organizations make informed business decisions, including new revenue opportunities, more effective marketing, better customer service, improved operational efficiency and competitive advantages over rivals. Numerous industrial and research databases have quality issues including outliers, noise, missing values, and so on. In fact, it is not uncommon to encounter databases that have up to a half of the entries missing, making it very difficult to mine them using data analysis methods that can work only with complete data. Currently, comprehensive analysis and research of quality standards and quality assessment methods for big data are lacking. First, this talk summarizes reviews of data quality research. We then analyze the data characteristics of the Big Data environment, and present quality challenges faced by Big Data. Finally, we construct a dynamic assessment process for data quality.
Description:
The HPC field is constantly expanding, increasing its geographical and scientific reach. This expansion brings more diversity, enriching the community, but also more training and educational challenges. To address the educational gaps and ensure a continuous growth of HPC practitioners a number of international initiatives appeared within the HPC educational community. The PRACE online courses (MOOCs), ACM SIGHPC Education Chapter, HPC Carpentry and International HPC Certification Program are some of them. For these initiatives to thrive and be useful they need input and support from community members around the world. Come, join us or give your feedback! It’s needed.
Target Audience:
Anyone involved or interested in HPC education and training, including students.
Outline of Programme:
This session will focus on presenting work done to help address educational challenges faced by the HPC community. More specifically, the session will focus on the PRACE online courses (MOOCs), ACM SIGHPC Education Chapter, HPC Carpentry and International HPC Certification Program.
Although many of the challenges faced by HPC educators are common and don’t depend on a specific location or affiliation, some such challenges are more specific to groups of users. Quite often accommodating the extra needs of a smaller group, not only benefits the group in question, but can also benefit the community as a whole. If done correctly, small changes to the education material can improve the learning experience of a significant number of leaners.
Besides presenting the ideas behind the above-mentioned educational efforts, the session will aim to understand the challenges and educational needs faced by the African HPC community to influence the direction in which the global HPC training and education efforts are heading.
The session is meant to be useful to all participants so it will include a number of brief presentations, followed by Q&A sessions and a general discussion. The speakers are either directly involved in or have links to all of the above mentioned initiatives so are in position to answer questions and act on the received feedback.
The proposed session outline:
Presentations interspaced with Q&As:
• MOOCs
• HPC Carpentry
• HPC Certification Program
• ACM SIGHPC Education chapter
Presentations and Q&A sessions will be followed by a discussion on interests, challenges, suggestions for improvements etc., possibly with a survey running in the background to drive the conversation.
The session will then close with a summary of its outcomes, any actions identified during the session and information for participants on how to continue the conversation and engagement.
That research data should be shared with the rest of the world has become almost evangelical in nature. This paper will try to answer the following questions:
• What are the (real) reasons for ‘forcing’ scientists to open their data, even if they are not ready to do so?
• What right have non-scientists (and scientists) to push indiscriminately for the sharing of data without taking the nuances of research into consideration?
This paper will look at the physical characteristics of research data before it can be shared, a case study of the public humiliation of a excellent researcher who did not wish to share his data, as well as the advantages and disadvantages of sharing research data.
For the promise of genomics to drive precision medicine to be realized, numerous human genome sequence samples and high depth coverage and from a variety of populations is required. Through initiatives like H3Africa, population scale genomics is being applied to help characterize diverse African populations that have currently not been featured in the genomics revolution. African genomes represent an uncharacterized portion of human genomics and may hold opportunities for novel discoveries.
Genomics data however is big data, the collection, transfer, storage and analyses of genomics data poses challenges to the infrastructure needed to manage such data. Key to human genomics data is governance policies that control the access of sensitive genetic data and achieving a balance between not being a barrier to access data that drives scientific discoveries, and at the same time protect the individuals that provide data. Associated with human genomics data is the rich diversity of meta-data that provide contextual information of the data. Some of the meta-data can be shared, other meta-data that might possibly make a sample identifiable will need to be under controlled access. Meta-data standards that are interoperable are required for the sharing of data and also for any meta-analysis to be conducted and form the core of good data management when dealing with copious volumes of data. The adoption of FAIR principals (Findable, Accessible, Interoperable and Reusable) is becoming prevalent across all scientific disciplines and in some cases mandated by different funding agencies. The technological adoption needed to turn FAIR principals into reality is non-trivial, but achievable and should be incorporated within good data management practices.
For genomics data to be translated into actionable outcomes with impacts, intensive computational processing and analyses of the data needs to be undertaken to convert information into knowledge. Traditionally in bioinformatics numerous software packages, versions and dependencies are needed to undertake an analyses which can run into 100s or 1000s of computational hours making reproducibility of the science a key issue. Development of computational workflows and containerized environments with user defined software stacks which can be hosted on various high performance computing systems enables the sharing of software environments, versions, and computational workflows leading to reproducible science.
According to Feeney, Mary & Ross, Seamus in Information Technology in Humanities Scholarship, British Achievements, Prospects, and Barriers https://www.jstor.org/stable/20755828, The growth of digital media means that the main areas of scholarship can each benefit from expansions in their own way thanks to the infinite shareability of digital content.
With the ever-expanding affordances that new digital media and research tools provide us, the underlying question is what form of control, training, validation and systems need to be in-place to ensure efficient data management throughout the whole research life-cycle? Often organisations focus on the efficient data management during a specific phase; however, we need to ensure that best practices on research data management are applied throughout. Digital Scholarship Centres can play a key role in providing guidance, training and even systems to bring together an overarching portfolio of services and products that can steer digital and data intensive research.
This presentation will provide you with insight into the progress the University of Pretoria has made in establishing its own Digital Scholarship Centre, its successes and failures.
The drivers and principles for managing research data at the University of South Africa is driven by the increasing number of policies published by funders of research to ensure the validation of research results. The funder requirements expects the following from research outputs:
• To make provision for data reuse,
• To enable actionable and socially beneficial science from publicly funded research.
The purpose of the project is to ensure that research data is stored, preserved, retained, made accessible for use and reuse, and/or disposed of, according to legal, statutory, ethical and funding bodies’ requirements.
Unisa Library was mandated to investigate the Data Curation readiness at Unisa in 2010. Research was conducted and it was concluded that, that Data Curation cannot take place in isolation in the library. It was then decided to change the concept documents from data curation to Research Data Management, holistically. The project is driven from Unisa Library and in particular Information Resources Content Management (IRCM).
The paper will provide a case study analysis of the transition in the implementation of the Research Data Management and the interdependencies.
Research data management (or RDM) is a growing phenomenon, and is largely concerned with the creation, organization, storage, safeguarding, and dissemination of data related to research works. One of the key components of RDM activity is safeguarding the data, which undoubtedly includes security of the data. However, research data security is hardly discussed in the literature resulting to huge gap in assuring its security properties. Unarguably, research data comes in various forms and formats, and of different intrinsic value. It means that the sensitivity of research data may vary from research thematic area or environment, and may contain significant sensitive data that can necessitate some degree of security. This requires understanding of security requirements to proportionately safeguard research data. In this presentation, data security is examined in the context of research data especially confidentiality, integrity, authenticity, availability, non-repudiation and trustworthiness attributes. A framework is proposed on how researchers and research data management operators can preserve research data security proportionally.
Genetic algorithms and genetic progamming are metaheuristics taking inspi-
ration from Darwin’s theory of evolution to solve optimization problems. Ge-
netic algorithms explore a solution space to find solutions to problems while
genetic programming works in a program space to identify a program which
when executed will find an optimal solution to the problem. Both these
approaches have high runtimes when applied to complex problems and are
usually implemented using distributed computing in these instances. More
recently, genetic algorithms and genetic programming have been employed
by hyper-heuristics and have been used for the automated design of machine
learning and search techniques. Hyper-heuristics explore the heuristic space
rather than the solution space and hence search in the heuristic space is
mapped to the solution space. Automated design of machine learning and
search techniques is an emerging field aimed at removing the load of de-
sign, which is a time consuming task, from the researcher. This will also
enable non-experts to use machine learning toolkits that automate the de-
sign and hence allow the researcher to focus on the problem being solved.
The use of genetic algorithms and genetic programming in hyper-heuristics
and for automated design require additional processing time. The talk will
firstly look at the high performance computing architectures implemented
by our research group to reduce the runtimes of genetic programming and
genetic algorithms, particularly for hyper-heuristics and automated design.
An overview of some of the real-world problems that this has enabled us
to solve will then be presented. These include inducing human competitive
heuristics for solving timetabling problems, network intrusion detection in
the area of computer security, the automated design of techniques for finan-
cial forecasting, computer security, packing and logistics problems and the
introduction of multi-space search algorithms that perform search over more
than once space with applications in packing and forecasting.
Computational Fluid Dynamics (CFD) is a mathematical tool used for commercial and research problems, to predict flow- and thermal characteristics in a wide variety of applications. Aerotherm has been using CFD to improve and design flow-related systems in the South African industry since 1995. The ever increasing demand for higher performance drives the need for more comprehensive CFD models. Such high fidelity models, however, require sufficient resources to be completed in reasonable timeframes. This talk will discuss how the CHPC plays a crucial role in the success of key CFD projects at Aerotherm.
Computational material science is taking root in Kenya. Development of big data infrastructure, high-performance computing facilities and use of open-source-software will accelerate the growth of this field, develop competent manpower and improve the economy of the country. The role played by major public universities in Kenya such as Masinde Muliro University of Science and Technology, University of Eldoret, Kenyatta University and the University of Nairobi in tackling health, water, and energy-related issues using computational tools requires long-term funding support. The glaring gap in technology transfer from universities to industries in Kenya needs to be addressed in order to harness the full potential of computational material science.
The South African Weather Service (SAWS), as a national meteorological service provider for South Africa (SA) and Southern African Development Community (SADC) communities, aims to provide useful and innovative weather, climate, and related products and services. Part of SAWS services and products are derived from Numerical Weather Prediction (NWP) models, of which Unified model (UM) from the UK Metoffice is the lead model. Weather forecast models perform differently across different parts of the globe and for different weather phenomena. In order to ensure continuous delivery of high quality weather forecasts to SAWS stakeholders, other NWP models are examined. The Consortium for Small-scale Modelling (COSMO) model is evaluated and analysed in order to investigate whether it is more/less suitable for predicting high impact weather events over SA.
The COSMO model is a European limited area model driven from Icosahedral Non-hydrostatic (ICON) global model. The ICON model has a grid spacing of 13km globally, which allows the COSMO to be simulated at a fine horizontal resolution of less than 5km. In this study, the COSMO model is being run with a grid spacing of 4.4km and 40 vertical levels, which allows for accurate numerical prediction of near-surface weather conditions (e.g. clouds, fog, frontal precipitation) and simulation of severe weather events triggered by deep moist convection (supercell thunderstorms, intense mesoscale convective complexes, prefrontal squallline storms and heavy snowfall from wintertime mesocyclones.
Simulations of five high impact weather events over SA were done on the CHPC system. The COSMO model is run with a grid spacing of 4.4km (0.036°) over SADC with a domain ranging from 5 – 56 °E and 40 – 5 °S (1276 x 1026 grid points), using a timestep of 30 seconds and a 30 hour lead-time. The model is run on a large queue using 1728 cores with using default science settings. Both the pre-processing and model run completed in +/- 2.5 hours and the model output was in hourly intervals for 86 variables (80 in sigma coordinates and 6 in pressure coordinates).
The selected variables (parameters) from model output are then evaluated against ground observations from SAWS network, radar data, satellite observations, Global Precipitation Measurement (GPM) and National Centers for Environmental Prediction (NCEP) Reanalysis data. Selected parameters for evaluation include wind speed, surface temperature, total precipitation, cloud cover, mean sea-level pressure, surface pressure and geopotential height. The outcome of the model evaluation will be discussed in detail at the conference.
A brief overview of the Technology stack used by DIRISA to provide storage infrastructure and services will be provided.
Extended Title: An Investigation into the integration of Digital Object Identifier (DOI) and Data Upload Service for use in Data Intensive Research in South Africa (DIRISA).
Abstract: The use of Digital Object Identifiers has become an increasingly important topic within the research and data space. This is mainly due to the fact that researchers want their data to be easily findable, accessible and searchable. The Data Intensive Research Initiative of South Africa (DIRISA) has realised this need and is creating a system that can allow researchers to mint their own identifiers based on handle.net system. These identifiers can be referred to as Digital Object Identifiers (DOI’s). However, in order to mint identifiers the researchers need to upload data and provide metadata.
Research funders around the world are realizing the importance of data management planning. Increasingly research public funders globally, now require data management plans to be submitted with research proposals as a condition of funding support. This is also true in South Africa where research funders such as the National Research Foundation (NRF) already require researchers to provide a data management plan as part of their application for research funding. A data management plan can be described as a formal document that outlines how data will be dealt with by the researcher before, during and after the research has been conducted.
The DIRISA Data Management Planning (DMP) Tool has been developed to assist South African Researchers to create data management plans that align with the requirements of funders such as the National Research Foundation (NRF). This presentation will give a brief demo of the tool.
Over the past fifteen years or so, the number of modellers at Johnson Matthey has grown from one, to close to twenty; a firm indication of the importance that JM places in both atomistic, as well as meso/macroscale modelling. DFT modelling in particular, forms an integral part of JM’s research and development efforts in a wide range of areas, including catalysis, diagnostic services, pharmaceutical applications, and new battery and fuel cell technologies. The presentation will provide a brief overview of modelling at JM, highlighting the types of calculations that we do (with some real world examples), how modelling fits in with our experimental research programmes, the HPC resources that we utilize, and our interactions with the CHPC.
Acid deposition has been studied extensively over the Highveld region of South Africa due to its high density of acid emissions arising mostly from power generation and petrochemical production. Although deposition of acid species has been modelled using dispersion models, the modelling of atmospheric chemistry and acid deposition has not been undertaken to any meaningful extent in South Africa. Furthermore, the potential of WRFChem, the Weather, Research and Forecasting (WRF) model coupled with Chemistry, has yet to be realized within an African context mainly due to the high computing power required to compute even the smallest of domain sizes. With a domain size of over 300 000 km2, which equates to over 75 000 grid cells, over a two-year modelling period, computational power was a limiting factor to the success of the research. Access to high performance computing and more specifically the ability to split the processing across multiple processors is instrumental in running this computationally intensive model. And has been the limiting factor for further research in South Africa on this topic. This presentation provides a review of the benefits of using High Performance Computing for running WRFChem when compared with local micro-servers.
The national rail policy as formulated by the South African government requires that by the year 2050 all current narrow or Cape gauge main rail lines be replaced by standard gauge tracks. In addition, they require that a number of key role players, of which Transnet is but one, should work toward the implementation of high speed rail on these new tracks by the same year. The document further stipulates that the onus of responsibility of planning and developing the necessary skill in order to achieve the aforementioned, lies with these role players.
The first step that Transnet is taking to bridge this gap and move South Africa forward is called the MC25 project; a medium speed passenger commuter that will connect Gauteng and Polokwane. The aim of the project is both economic upliftment as well as a means of addressing a skills deficit with Transnet. The work that was completed was meant to address the latter, with specific focus on the external aerodynamics of a high speed train.
This project not only included a study of the external flow field surrounding a high speed train, but also the simultaneous optimization of the train nose and tail for drag and crosswind stability which were determined by the aerodynamic investigation to be the primary causes of concern for high speed trains. Due to the flow complexities and sheer magnitude of the simulations, the computing capabilities that the CHPC offered was critical to arriving at an optimal solution.
Functional genomics determines the biological functions of genes on a global scale by using large volumes of data obtained through techniques including next-generation sequencing (NGS). The application of NGS in biomedical research is gaining in momentum, and with its adoption becoming more widespread, there is an increasing need for access to customizable computational workflows that can simplify, and offer access to, computer-intensive analyses of genomic data. In this study, the Galaxy and Ruffus frameworks were designed and implemented with a view to addressing the challenges faced in biomedical research. Galaxy, a graphical web-based framework, allows researchers to build a graphical NGS data analysis pipeline for accessible, reproducible, and collaborative data-sharing. Ruffus, a UNIX command-line framework used by bioinformaticians as Python library to write scripts in an object-oriented style, allows for building a workflow in terms of task dependencies and execution logic. In this study, a dual data analysis technique was explored which focuses on a comparative evaluation of Galaxy and Ruffus frameworks that are used in composing analysis pipelines. To this end, we developed an analysis pipeline in Galaxy, and Ruffus, for the analysis of Mycobacterium tuberculosis sequence data. Furthermore, this study aimed to compare the Galaxy framework to Ruffus with preliminary analysis revealing that the analysis pipeline in Galaxy displayed a higher percentage of load and store instructions. In comparison, pipelines in Ruffus tended to be CPU bound and memory intensive. The CPU usage, memory utilization, and runtime execution are graphically represented in this study. Our evaluation suggests that workflow frameworks have distinctly different features from an ease of use, flexibility, and portability, to architectural designs. Therefore, in this CHPC Conference, I will discuss how we composed the NGS bioinformatics data analysis pipeline in the Galaxy and Ruffus workflow framework and the use of each framework in the analysis of Mycobacterium tuberculosis sequence data.
Delivering HPC solutions via cloud-based resources is still a technical challenge. Beside either running HPC workload on on-premise resources or entirely on cloud resources hybrid approaches can be used for providing a flexible and cost-effective way of running HPC workloads. Based on two examples, a turn-key SaaS solution (HyperworkUnlimited Virtual Appliance) and a Cloud Bursting scenario with PBS Control we will discuss what is involved in such setups and how to make it work smoothly.
With all the advances in massively parallel and multi-core computing with CPUs and accelerators, it is often overlooked whether the computational work is being done in an efficient manner. This efficiency is largely being determined at the application level and therefore puts the responsibility of sustaining a certain performance trajectory into the hands of the user. It is observed that the adoption rate of new hardware capabilities is decreasing and lead to a feeling of diminishing returns. At the same time, the well-known laws of parallel performance are limiting the perspective of a system builder. The presentation tries gives an overview of these challenges and what can be done to overcome them. The overview will be amended by a few case studies and optimization strategies on real applications.
All CHPC Principal Investigators (PI's) and Users (and prospective new PI's and users) are encouraged to attend this Birds-of-a-Feather (BoF) session.
Werner Janse van Rensburg will give an overview of usage patterns, research programme distributions and overall usage statistics of the CHPC Lengau cluster over the past 12 months.
In addition, discussions will focus on aspects such as:
Registration procedures, user accounts and policies
Storage (Lustre and Home)
Job Queues
Allocation Evaluation and CPU Hour Allocations
New GPU Cluster
Queries and Helpdesk
Etc.
This will be a good opportunity for attendees to ask questions and meet with CHPC staff members in an informal setting.
The HPC Ecosystems Project is continuing to support HPC deployment initiatives throughout Africa, with two new partner sites receiving their first HPC systems in 2018. Additionally, the Ecosystems Project continues to deliver educational content and training to address workforce preparation for both new users and seasoned professionals.
Although the CHPC is responsible for the leadership and oversight of the Project, the success of the Project greatly depends on the broader HPC Ecosystems community’s engagements and collaboration.
This CHPC-coordinated BoF will engage all stakeholders in helping to determine the roadmap for 2019 (and beyond). Particular focus will be given to discussing the following points:
The mandate of the South African Weather Service (SAWS) is to provide weather, climate and air quality related information, products, services and solutions that contribute to the safety of life and property in the air, land and sea. To deliver on its mandate, SAWS operates a comprehensive weather observation network and also runs models in operational mode. The observation network includes 14 operational some S-band and C-band radars which cover most of the country. SAWS also operates a lighting detection network which has been in operation since 2006. Further SAWS also access satellite information in near real time and this information is especially helpful for areas where there is no radar network. The United Kingdom Met Office (UKMO) Unified Model (UM) is used for short range forecasting and is operated with a grid spacing of 4.4km and 1.5 km. For the medium range timescale, SAWS downloads data from the European Centre for Medium-Range Weather Forecasts (ECMWF) and National Centers for Environmental Prediction (NCEP) Global Ensemble Forecasting System (GEFS). In this presentation the performance of SAWS in issuing warnings of high impact weather events that took place in 2017 will be assessed against the model simulations that are produced both at SAWS as well as the ECMWF simulations. Moreover, the assessment of the performance will also be checked against the availability of the radar and satellite data in the nowcasting timescale. The general statistics will be presented for floods and heavy rainfall events, thunderstorms including hail storms and tornadoes, as well as strong winds that were reported in print media. Initial results indicate that SAWS issued warnings for 21/26 veld fires, 28/51 heavy rain or flooding events, 18/32 thunderstorms and 9/18 strong winds 24 hours in advance. The data that is valuable for warnings that are issued 24 hours or longer before the event occurs is produced by models. Closer to the events, the critical infrastructure is that associated with near real-time observations. In the presentation we will also show the effects of the availability of the observations to warnings issued by SAWS. The presentation will also highlight challenges associated with managing both observation and model simulation data.
One of the principal benefits of and drivers for Open Science is the open use of shared infrastructures for science. The digital age offers huge opportunities for the accelerated creation of and networked access to data for research. The exploitation of this through the Internet, the Web, Grids, e-Science, Research Infrastructures, Open Science Clouds and Platforms has been a major driver of many areas of scientific progress and a key feature of science policy and investment (in particular disciplines or in cross-disciplinary programmes).
Global north investment in research infrastructures, data commons and Open Science Clouds is a direct result of the scientific benefits achieved in the last 15-20 years through networked access to data resources and computing power. It is essential that African countries and Africa as a continent keep pace with these developments.
There is growing awareness globally that it is not enough to create network or compute infrastructure in a way that is abstracted from research disciplines and concrete use cases. Moreover, and of crucial importance is the recognition that data require the investment of stewardship to ensure that they can be fully exploited. Again this has been variously expressed: data should be a first class object, data as infrastructure, intelligent openness or FAIR data. With ever more useful refinements, these arguments point towards the same end: a necessary feature of research infrastructures is a rich environment of data resources that are Findable, Accessible, Interoperable and Reusable (FAIR); as Open as Possible and stewarded for the long term as part of the record and resource of science.
The most obvious challenges, relate investment and the funds needed for such infrastructures. But they also relate to coordination. A great deal can be achieved through coordination of resources and investment in shared infrastructures. The development of high-speed educational research networks and improved access to HPC clusters is proceeding and there are some instances of experimentation with cloud computing for science. The challenge is to achieve the vision, strategy and coordination that will ensure these investments will enable to African researchers to engage with the Open, data-intensive science of the 21st century. This means building on the infrastructure a shared layer providing access to compute and research tools of various sorts and to data that is Open and FAIR. A lot can be learnt from activities globally, particularly in Europe and Australia. A major challenge is to determine how such levels of coordination can be achieved between African research systems and institutions. Without such coordination and the attendant economies of scale and cost sharing, the development of research infrastructures will be far slower, more expensive, ad hoc and susceptible to quicker technical redundancy or obsolescence.
This paper will report on an African Open Science Platform ICT Infrastructure Framework and Roadmap, guiding African countries towards preparing for a coordinated initiative, addressing Open Science and Open Access needs across disciplines. South Africa continues to be a main player as will be indicated in this paper, together with countries such as Kenya. This paper will demonstrate that collaboration and coordination are possible and much needed, towards addressing the objectives of the SA National Development Plan and the UN Science, Technology and Innovation Strategy for Africa 2024.
ILIFU is the isiXhosa word for Cloud, an apt representation of the first node in the data infrastructure funded by the Department of Science and Technology to support the National Integrated Cyberinfrastructure System of South Africa. This presentation will provide an overview of the ILIFU model of shared infrastructure as a data-intensive research facility for big data management, storage and analysis. It will focus primarily on the Research Data Management (RDM) component of the project from conception to its current level of operationalisation.
The current progress will be shared with the community to raise awareness of the RDM project, its methodology and its structure. The overall project has a strong emphasis on the development of policies and guidelines to assist researchers who will be working on the data intensive infrastructure. The project is also developing a Work Integrated Learning programme to enable the placement of postgraduate students and mid-career personnel with the various projects in order to assist with the implementation of the RDM policies and guidelines as well as gaining work experience in data science. Significantly, the RDM research project fulfils an important additional function in support of institutional capacity building across participating institutions, where project findings will serve to promote institutional policy and service development.
The presentation will offer insights into data governance, the challenges and opportunities of the ILIFU RDM project, and the interaction between scientists and information professionals in building an African data infrastructure.
Coreless Continuum Computer Architecture
Thomas Sterling (Department of Intelligent Systems Engineering, School of Informatics, Computing, and Engineering, Indiana University)
As introduced at CHPC-17, a new era is dawning: the Neo-Digital Age as semiconductor fabrication enters its final phase with nano-scale feature size. For the first time in three decades, fundamental technology limitations of device density, power consumption, clock rate, and instruction level parallelism are demanding revolutionary practices to further expand HPC capabilities into the trans-exascale performance regime and perhaps beyond. Among the domains of possible change are those uncovered in the prevailing assumptions embodied by the many generations exploiting Moore’s Law, which no longer apply. These include 1) the use of static methods for parallel application programming and static resource management, and 2) the implicit underlying principles buried in the von Neumann architecture of which essentially all HPC systems are since derivatives. An alternative strategy, bordering on a paradigm shift, reverses conventional methodologies a) to employ dynamic adaptive methods of resource management and task scheduling, and b) to eliminate von Neumann bottlenecks, eliminating even the basic core that has endured as the foundation of HPC architecture design since the late 1940s. Central to the forward looking revolution is the elimination of the FPU/ALU as the precious resource with the implied objective function of highest possible utilization. The Continuum Computer Architecture (CCA) has been proposed as a family of alternative architectures that address the current challenge and exploit the future opportunities by fully integration of logic, memory, control, and communication within a single highly replicated computing cell. First suggested an abstract in last year’s last Keynote, this presentation will expose key details in structure and semantics that will layout the conceptual scaffolding for future development and application such as machine learning and intelligence, planning, scheduling, hypothesis testing, facial recognition and many other graph-based application. Questions will be encouraged by participants throughout the presentation as well as the Q&A session at the end.
The CHPC will be engaging with its non-academic users on Thursday 6 December. The format of the meeting will be similar to previous years, and mostly involves receiving feedback from the active commercial and public sector user base.
The purpose of the meeting is to strengthen collaboration with the non-academic users, exchange experiences and discuss the way forward. Attendees who are currently active CHPC users will be requested to give a short introduction to their HPC-related work.
Although the focus of the meeting will be to engage with past and current non-academic users, the meeting will also be open to any individual or organisation wishing to learn more about what the CHPC offers to industry and the public sector.
The CHPC will be engaging with its non-academic users on Thursday 6 December. The format of the meeting will be similar to previous years, and mostly involves receiving feedback from the active commercial and public sector user base.
The purpose of the meeting is to strengthen collaboration with the non-academic users, exchange experiences and discuss the way forward. Attendees who are currently active CHPC users will be requested to give a short introduction to their HPC-related work.
Although the focus of the meeting will be to engage with past and current non-academic users, the meeting will also be open to any individual or organisation wishing to learn more about what the CHPC offers to industry and the public sector.
A dedicated Chemistry and Material Science SIG session consisting of keynotes, ordinary oral presentations, student oral presentations as well as flash poster presentations. The program will also include the annual SIG meeting.
Program:
09:00 Keynote: John Mack, Rhodes University, The use of Gaussian 09 to identify trends in the optical properties and electronic structures of porphyrin and BODIPY dyes
09:30 Talk: Marianne Conradie, UFS, Determination of the steric effect of a bulky R substituent on the reaction mechanism of [Rh(CH3COCRCOCH3)(CO)(PPh3)] + CH3I
09:50 Talk: Cliffton Masedi, UL, Computer Simulation and Phase Diagram Prediction of Li2S1-xSex Systems
10:10 Talk: Thishana Singh, DUT, Biosynthesis and computational analysis of amine-ended dual thiol ligand functionalized gold nanoparticles
10:30 Morning Tea Break
11:00 Student talk: Cameron Matthews, NMU, A theoretical study of electronic (QTAIM NBO) and geometric structures of a series of lanthanide trichloride LnCl3 complexes (Ln=La-Lu) with the N-donor ligand bis(2-pyridylmethyl)amine (DPA)
11:15 Student talk: Elkana Rugut, Wits, Thermoelectric properties of CdAl2O4 spinel
11:30 Student talk: Fisayo Olotu, UKZN, Delving through the wonders of green tea - Probing the dynamic selectivity of polyphenol Epigallocatechin gallate
11:45 Student talk: Francis Opoku, UJ, Tuning the electronic structures
12:00 Student poster flash presentation: Ramogohlo Diale, UL, Computational modelling of TiPd-Ir high temperature shape memory alloys
12:05 Student poster flash presentation: Sharlene-Asia Naicker, UKZN, Computational studies of the corrosion mechanisms in transformers
12:10 Student poster flash presentation: Surya Narayana Maddila, UKZN, Synthesis and spectral properties of 5-substituted 1H-tetrazole derivatives using with Mn/HAp catalyst and DFT calculations
12:15 Student poster flash presentation: Nkosinathi Malaza, Wits, Double-layer capacitance from ab initio simulations
12:20 Student poster flash presentation: Jonathan Gertzen, UCT, MAX Phases as an Electrocatalyst Support Material: A DFT Study
12:25 Discussion
12:30 Lunch Break
13:30 Talk: Nkululeko Damoyi, MUT, A DFT study of the ODH of n-hexane over isolated H3VO4 and H4V2O7
13:50 Student talk: Adebayo Adeniyi, UFS, The theoretical investigation of reduction potentials and and spectroscopic properties of nitrobenzene and keto-enol molecules
14:05 Student talk: Abdulgaffar, UP, Hybrid functional study of BiOi, BiCs, BiBsHi and BiOiOi Complexes in Silicon
14:20 Student talk: Joseph Simfukwe, UP, First Principles study of Zn/Cu doped hematite surfaces for Photoelectrochemical water splitting
14:35 Talk: Khalid Ahmed, UKZN, Understanding the influence of pH/pKa by theoretical and computational method for loading/release of Ibuprofen (To Be Confirmed)
14:55 Discussion
15:00 Afternoon Tea Break
15:30 SIG Meeting
HPC Carpentry is an introduction to working on the CHPC cluster with an overview of the concepts necessary to make effective use of a high performance computing (HPC) cluster along with practical introduction to submitting jobs, finding software and splitting up your data so that it can be processed in parallel. This workshop is appropriate for people who are new to the world of clusters and want a practical introduction to the benefits they can derive from HPC and high throughput computing (HTC).
Molecular Dynamics and analysis can be complicated for novitiates and researchers from neighbouring disciplines. Building upon the Galaxy Project platform, BRIDGE (Biomolecular Reaction & Interaction Dynamics Global Environment) is a web application that provides the ability to get started running molecular dynamics and analyses using curated workflows.
Target Audience: Computational chemists, biologists, protein scientists, crystallographers and anyone interested in running and analysing molecular dynamics simulations who is not familiar with the command line.
Prerequisites: Basic understanding of chemistry. A basic understanding of or keen interest in molecular dynamics.
Special requirements: Bring a laptop. Have access to the CHPC cluster. Register with BRIDGE. Install molecular viewer e.g. VMD.
Day: Thursday (6 Dec)
Duration: 1 day
Size: 15 seats
Description:
Join us for one day of hands-on sessions on Artificial Intelligence, and Machine & Deep Learning. Experience a unique opportunity to test out the latest performance optimized frameworks and tools, advanced coding knowledge and best practices to get started implementing AI guided by experts from Intel®.
Target Audience:
Data Scientist, application developers and HPC benchmarkers targeting the deep learning and machine learning domain.
Prerequisites:
Beginning to intermediate level of domain AI knowledge.
Basic skills of programming , ideally some Python knowldege
Type of tutorial: Mix of lectures and hands-on tutorials
Special Requirements:
Attendees should bring their laptop with an SSH- & VNC client
Attendees will get for hands on-labs also access to the CHPC cluster
Outline of full syllabus:
08:00 Registration
09:00 Introduction
• Introduction round & Agenda
• Introduction Intel Software Developer Tools
• Introduction to Machine Learning / Deep Learning
10:30 Morning Refreshment Break
11:00
Classic Machine Learning Tools
• Intel performance Libraries - MKL & DAAL
• Intel Distribution for Python (IDP) Introduction
• IDP Hands on labs
o NumPy & MKL
o K-Means Clustering & DAAL
o SVM & DAAL
12:30 Lunch
13:30
Deep Learning (DL) Tools
• Intel performance Libraries for DL – MKL-DNN & MLSL
• Intel optimized Frameworks / TensorFlow
o TensorFlow Image Classification Hands-on Lab
o Introduction simple CNN
o Monitored Training Session
15:00 Afternoon Refreshment Break
15:30
Deep Learning Tools (cont’d)
• Intel optimized Frameworks / TensorFlow (cnt’d)
o Horovod – distributed classification
o Importing external Images
o Custom batches
• Benchmarking distributed TensorFlow (BKMs)
• Deep Learning Scaling – large scale results (BigDL/Spark)
• Wrap-Up & Q&A
17:00 End of Day
Description:
Join us for one day of hands-on sessions on Artificial Intelligence, and Machine & Deep Learning. Experience a unique opportunity to test out the latest performance optimized frameworks and tools, advanced coding knowledge and best practices to get started implementing AI guided by experts from Intel®.
Target Audience:
Data Scientist, application developers and HPC benchmarkers targeting the deep learning and machine learning domain.
Prerequisites:
Beginning to intermediate level of domain AI knowledge.
Basic skills of programming , ideally some Python knowldege
Type of tutorial: Mix of lectures and hands-on tutorials
Special Requirements:
Attendees should bring their laptop with an SSH- & VNC client
Attendees will get for hands on-labs also access to the CHPC cluster
Outline of full syllabus:
08:00 Registration
09:00 Introduction
• Introduction round & Agenda
• Introduction Intel Software Developer Tools
• Introduction to Machine Learning / Deep Learning
10:30 Morning Refreshment Break
11:00
Classic Machine Learning Tools
• Intel performance Libraries - MKL & DAAL
• Intel Distribution for Python (IDP) Introduction
• IDP Hands on labs
o NumPy & MKL
o K-Means Clustering & DAAL
o SVM & DAAL
12:30 Lunch
13:30
Deep Learning (DL) Tools
• Intel performance Libraries for DL – MKL-DNN & MLSL
• Intel optimized Frameworks / TensorFlow
o TensorFlow Image Classification Hands-on Lab
o Introduction simple CNN
o Monitored Training Session
15:00 Afternoon Refreshment Break
15:30
Deep Learning Tools (cont’d)
• Intel optimized Frameworks / TensorFlow (cnt’d)
o Horovod – distributed classification
o Importing external Images
o Custom batches
• Benchmarking distributed TensorFlow (BKMs)
• Deep Learning Scaling – large scale results (BigDL/Spark)
• Wrap-Up & Q&A
17:00 End of Day
The CHPC will be engaging with its non-academic users on Thursday 6 December. The format of the meeting will be similar to previous years, and mostly involves receiving feedback from the active commercial and public sector user base.
The purpose of the meeting is to strengthen collaboration with the non-academic users, exchange experiences and discuss the way forward. Attendees who are currently active CHPC users will be requested to give a short introduction to their HPC-related work.
Although the focus of the meeting will be to engage with past and current non-academic users, the meeting will also be open to any individual or organisation wishing to learn more about what the CHPC offers to industry and the public sector.
The CHPC will be engaging with its non-academic users on Thursday 6 December. The format of the meeting will be similar to previous years, and mostly involves receiving feedback from the active commercial and public sector user base.
The purpose of the meeting is to strengthen collaboration with the non-academic users, exchange experiences and discuss the way forward. Attendees who are currently active CHPC users will be requested to give a short introduction to their HPC-related work.
Although the focus of the meeting will be to engage with past and current non-academic users, the meeting will also be open to any individual or organisation wishing to learn more about what the CHPC offers to industry and the public sector.
A dedicated Chemistry and Material Science SIG session consisting of keynotes, ordinary oral presentations, student oral presentations as well as flash poster presentations. The program will also include the annual SIG meeting.
Program:
09:00 Keynote: John Mack, Rhodes University, The use of Gaussian 09 to identify trends in the optical properties and electronic structures of porphyrin and BODIPY dyes
09:30 Talk: Marianne Conradie, UFS, Determination of the steric effect of a bulky R substituent on the reaction mechanism of [Rh(CH3COCRCOCH3)(CO)(PPh3)] + CH3I
09:50 Talk: Cliffton Masedi, UL, Computer Simulation and Phase Diagram Prediction of Li2S1-xSex Systems
10:10 Talk: Thishana Singh, DUT, Biosynthesis and computational analysis of amine-ended dual thiol ligand functionalized gold nanoparticles
10:30 Morning Tea Break
11:00 Student talk: Cameron Matthews, NMU, A theoretical study of electronic (QTAIM NBO) and geometric structures of a series of lanthanide trichloride LnCl3 complexes (Ln=La-Lu) with the N-donor ligand bis(2-pyridylmethyl)amine (DPA)
11:15 Student talk: Elkana Rugut, Wits, Thermoelectric properties of CdAl2O4 spinel
11:30 Student talk: Fisayo Olotu, UKZN, Delving through the wonders of green tea - Probing the dynamic selectivity of polyphenol Epigallocatechin gallate
11:45 Student talk: Francis Opoku, UJ, Tuning the electronic structures
12:00 Student poster flash presentation: Ramogohlo Diale, UL, Computational modelling of TiPd-Ir high temperature shape memory alloys
12:05 Student poster flash presentation: Sharlene-Asia Naicker, UKZN, Computational studies of the corrosion mechanisms in transformers
12:10 Student poster flash presentation: Surya Narayana Maddila, UKZN, Synthesis and spectral properties of 5-substituted 1H-tetrazole derivatives using with Mn/HAp catalyst and DFT calculations
12:15 Student poster flash presentation: Nkosinathi Malaza, Wits, Double-layer capacitance from ab initio simulations
12:20 Student poster flash presentation: Jonathan Gertzen, UCT, MAX Phases as an Electrocatalyst Support Material: A DFT Study
12:25 Discussion
12:30 Lunch Break
13:30 Talk: Nkululeko Damoyi, MUT, A DFT study of the ODH of n-hexane over isolated H3VO4 and H4V2O7
13:50 Student talk: Adebayo Adeniyi, UFS, The theoretical investigation of reduction potentials and and spectroscopic properties of nitrobenzene and keto-enol molecules
14:05 Student talk: Abdulgaffar, UP, Hybrid functional study of BiOi, BiCs, BiBsHi and BiOiOi Complexes in Silicon
14:20 Student talk: Joseph Simfukwe, UP, First Principles study of Zn/Cu doped hematite surfaces for Photoelectrochemical water splitting
14:35 Talk: Khalid Ahmed, UKZN, Understanding the influence of pH/pKa by theoretical and computational method for loading/release of Ibuprofen (To Be Confirmed)
14:55 Discussion
15:00 Afternoon Tea Break
15:30 SIG Meeting
HPC Carpentry is an introduction to working on the CHPC cluster with an overview of the concepts necessary to make effective use of a high performance computing (HPC) cluster along with practical introduction to submitting jobs, finding software and splitting up your data so that it can be processed in parallel. This workshop is appropriate for people who are new to the world of clusters and want a practical introduction to the benefits they can derive from HPC and high throughput computing (HTC).
Molecular Dynamics and analysis can be complicated for novitiates and researchers from neighbouring disciplines. Building upon the Galaxy Project platform, BRIDGE (Biomolecular Reaction & Interaction Dynamics Global Environment) is a web application that provides the ability to get started running molecular dynamics and analyses using curated workflows.
Target Audience: Computational chemists, biologists, protein scientists, crystallographers and anyone interested in running and analysing molecular dynamics simulations who is not familiar with the command line.
Prerequisites: Basic understanding of chemistry. A basic understanding of or keen interest in molecular dynamics.
Special requirements: Bring a laptop. Have access to the CHPC cluster. Register with BRIDGE. Install molecular viewer e.g. VMD.
Day: Thursday (6 Dec)
Duration: 1 day
Size: 15 seats
Description:
Join us for one day of hands-on sessions on Artificial Intelligence, and Machine & Deep Learning. Experience a unique opportunity to test out the latest performance optimized frameworks and tools, advanced coding knowledge and best practices to get started implementing AI guided by experts from Intel®.
Target Audience:
Data Scientist, application developers and HPC benchmarkers targeting the deep learning and machine learning domain.
Prerequisites:
Beginning to intermediate level of domain AI knowledge.
Basic skills of programming , ideally some Python knowldege
Type of tutorial: Mix of lectures and hands-on tutorials
Special Requirements:
Attendees should bring their laptop with an SSH- & VNC client
Attendees will get for hands on-labs also access to the CHPC cluster
Outline of full syllabus:
08:00 Registration
09:00 Introduction
• Introduction round & Agenda
• Introduction Intel Software Developer Tools
• Introduction to Machine Learning / Deep Learning
10:30 Morning Refreshment Break
11:00
Classic Machine Learning Tools
• Intel performance Libraries - MKL & DAAL
• Intel Distribution for Python (IDP) Introduction
• IDP Hands on labs
o NumPy & MKL
o K-Means Clustering & DAAL
o SVM & DAAL
12:30 Lunch
13:30
Deep Learning (DL) Tools
• Intel performance Libraries for DL – MKL-DNN & MLSL
• Intel optimized Frameworks / TensorFlow
o TensorFlow Image Classification Hands-on Lab
o Introduction simple CNN
o Monitored Training Session
15:00 Afternoon Refreshment Break
15:30
Deep Learning Tools (cont’d)
• Intel optimized Frameworks / TensorFlow (cnt’d)
o Horovod – distributed classification
o Importing external Images
o Custom batches
• Benchmarking distributed TensorFlow (BKMs)
• Deep Learning Scaling – large scale results (BigDL/Spark)
• Wrap-Up & Q&A
17:00 End of Day
Description:
Join us for one day of hands-on sessions on Artificial Intelligence, and Machine & Deep Learning. Experience a unique opportunity to test out the latest performance optimized frameworks and tools, advanced coding knowledge and best practices to get started implementing AI guided by experts from Intel®.
Target Audience:
Data Scientist, application developers and HPC benchmarkers targeting the deep learning and machine learning domain.
Prerequisites:
Beginning to intermediate level of domain AI knowledge.
Basic skills of programming , ideally some Python knowldege
Type of tutorial: Mix of lectures and hands-on tutorials
Special Requirements:
Attendees should bring their laptop with an SSH- & VNC client
Attendees will get for hands on-labs also access to the CHPC cluster
Outline of full syllabus:
08:00 Registration
09:00 Introduction
• Introduction round & Agenda
• Introduction Intel Software Developer Tools
• Introduction to Machine Learning / Deep Learning
10:30 Morning Refreshment Break
11:00
Classic Machine Learning Tools
• Intel performance Libraries - MKL & DAAL
• Intel Distribution for Python (IDP) Introduction
• IDP Hands on labs
o NumPy & MKL
o K-Means Clustering & DAAL
o SVM & DAAL
12:30 Lunch
13:30
Deep Learning (DL) Tools
• Intel performance Libraries for DL – MKL-DNN & MLSL
• Intel optimized Frameworks / TensorFlow
o TensorFlow Image Classification Hands-on Lab
o Introduction simple CNN
o Monitored Training Session
15:00 Afternoon Refreshment Break
15:30
Deep Learning Tools (cont’d)
• Intel optimized Frameworks / TensorFlow (cnt’d)
o Horovod – distributed classification
o Importing external Images
o Custom batches
• Benchmarking distributed TensorFlow (BKMs)
• Deep Learning Scaling – large scale results (BigDL/Spark)
• Wrap-Up & Q&A
17:00 End of Day
The CHPC will be engaging with its non-academic users on Thursday 6 December. The format of the meeting will be similar to previous years, and mostly involves receiving feedback from the active commercial and public sector user base.
The purpose of the meeting is to strengthen collaboration with the non-academic users, exchange experiences and discuss the way forward. Attendees who are currently active CHPC users will be requested to give a short introduction to their HPC-related work.
Although the focus of the meeting will be to engage with past and current non-academic users, the meeting will also be open to any individual or organisation wishing to learn more about what the CHPC offers to industry and the public sector.
The CHPC will be engaging with its non-academic users on Thursday 6 December. The format of the meeting will be similar to previous years, and mostly involves receiving feedback from the active commercial and public sector user base.
The purpose of the meeting is to strengthen collaboration with the non-academic users, exchange experiences and discuss the way forward. Attendees who are currently active CHPC users will be requested to give a short introduction to their HPC-related work.
Although the focus of the meeting will be to engage with past and current non-academic users, the meeting will also be open to any individual or organisation wishing to learn more about what the CHPC offers to industry and the public sector.
A dedicated Chemistry and Material Science SIG session consisting of keynotes, ordinary oral presentations, student oral presentations as well as flash poster presentations. The program will also include the annual SIG meeting.
Program:
09:00 Keynote: John Mack, Rhodes University, The use of Gaussian 09 to identify trends in the optical properties and electronic structures of porphyrin and BODIPY dyes
09:30 Talk: Marianne Conradie, UFS, Determination of the steric effect of a bulky R substituent on the reaction mechanism of [Rh(CH3COCRCOCH3)(CO)(PPh3)] + CH3I
09:50 Talk: Cliffton Masedi, UL, Computer Simulation and Phase Diagram Prediction of Li2S1-xSex Systems
10:10 Talk: Thishana Singh, DUT, Biosynthesis and computational analysis of amine-ended dual thiol ligand functionalized gold nanoparticles
10:30 Morning Tea Break
11:00 Student talk: Cameron Matthews, NMU, A theoretical study of electronic (QTAIM NBO) and geometric structures of a series of lanthanide trichloride LnCl3 complexes (Ln=La-Lu) with the N-donor ligand bis(2-pyridylmethyl)amine (DPA)
11:15 Student talk: Elkana Rugut, Wits, Thermoelectric properties of CdAl2O4 spinel
11:30 Student talk: Fisayo Olotu, UKZN, Delving through the wonders of green tea - Probing the dynamic selectivity of polyphenol Epigallocatechin gallate
11:45 Student talk: Francis Opoku, UJ, Tuning the electronic structures
12:00 Student poster flash presentation: Ramogohlo Diale, UL, Computational modelling of TiPd-Ir high temperature shape memory alloys
12:05 Student poster flash presentation: Sharlene-Asia Naicker, UKZN, Computational studies of the corrosion mechanisms in transformers
12:10 Student poster flash presentation: Surya Narayana Maddila, UKZN, Synthesis and spectral properties of 5-substituted 1H-tetrazole derivatives using with Mn/HAp catalyst and DFT calculations
12:15 Student poster flash presentation: Nkosinathi Malaza, Wits, Double-layer capacitance from ab initio simulations
12:20 Student poster flash presentation: Jonathan Gertzen, UCT, MAX Phases as an Electrocatalyst Support Material: A DFT Study
12:25 Discussion
12:30 Lunch Break
13:30 Talk: Nkululeko Damoyi, MUT, A DFT study of the ODH of n-hexane over isolated H3VO4 and H4V2O7
13:50 Student talk: Adebayo Adeniyi, UFS, The theoretical investigation of reduction potentials and and spectroscopic properties of nitrobenzene and keto-enol molecules
14:05 Student talk: Abdulgaffar, UP, Hybrid functional study of BiOi, BiCs, BiBsHi and BiOiOi Complexes in Silicon
14:20 Student talk: Joseph Simfukwe, UP, First Principles study of Zn/Cu doped hematite surfaces for Photoelectrochemical water splitting
14:35 Talk: Khalid Ahmed, UKZN, Understanding the influence of pH/pKa by theoretical and computational method for loading/release of Ibuprofen (To Be Confirmed)
14:55 Discussion
15:00 Afternoon Tea Break
15:30 SIG Meeting
HPC Carpentry is an introduction to working on the CHPC cluster with an overview of the concepts necessary to make effective use of a high performance computing (HPC) cluster along with practical introduction to submitting jobs, finding software and splitting up your data so that it can be processed in parallel. This workshop is appropriate for people who are new to the world of clusters and want a practical introduction to the benefits they can derive from HPC and high throughput computing (HTC).
Molecular Dynamics and analysis can be complicated for novitiates and researchers from neighbouring disciplines. Building upon the Galaxy Project platform, BRIDGE (Biomolecular Reaction & Interaction Dynamics Global Environment) is a web application that provides the ability to get started running molecular dynamics and analyses using curated workflows.
Target Audience: Computational chemists, biologists, protein scientists, crystallographers and anyone interested in running and analysing molecular dynamics simulations who is not familiar with the command line.
Prerequisites: Basic understanding of chemistry. A basic understanding of or keen interest in molecular dynamics.
Special requirements: Bring a laptop. Have access to the CHPC cluster. Register with BRIDGE. Install molecular viewer e.g. VMD.
Day: Thursday (6 Dec)
Duration: 1 day
Size: 15 seats
Description:
Join us for one day of hands-on sessions on Artificial Intelligence, and Machine & Deep Learning. Experience a unique opportunity to test out the latest performance optimized frameworks and tools, advanced coding knowledge and best practices to get started implementing AI guided by experts from Intel®.
Target Audience:
Data Scientist, application developers and HPC benchmarkers targeting the deep learning and machine learning domain.
Prerequisites:
Beginning to intermediate level of domain AI knowledge.
Basic skills of programming , ideally some Python knowldege
Type of tutorial: Mix of lectures and hands-on tutorials
Special Requirements:
Attendees should bring their laptop with an SSH- & VNC client
Attendees will get for hands on-labs also access to the CHPC cluster
Outline of full syllabus:
08:00 Registration
09:00 Introduction
• Introduction round & Agenda
• Introduction Intel Software Developer Tools
• Introduction to Machine Learning / Deep Learning
10:30 Morning Refreshment Break
11:00
Classic Machine Learning Tools
• Intel performance Libraries - MKL & DAAL
• Intel Distribution for Python (IDP) Introduction
• IDP Hands on labs
o NumPy & MKL
o K-Means Clustering & DAAL
o SVM & DAAL
12:30 Lunch
13:30
Deep Learning (DL) Tools
• Intel performance Libraries for DL – MKL-DNN & MLSL
• Intel optimized Frameworks / TensorFlow
o TensorFlow Image Classification Hands-on Lab
o Introduction simple CNN
o Monitored Training Session
15:00 Afternoon Refreshment Break
15:30
Deep Learning Tools (cont’d)
• Intel optimized Frameworks / TensorFlow (cnt’d)
o Horovod – distributed classification
o Importing external Images
o Custom batches
• Benchmarking distributed TensorFlow (BKMs)
• Deep Learning Scaling – large scale results (BigDL/Spark)
• Wrap-Up & Q&A
17:00 End of Day
Description:
Join us for one day of hands-on sessions on Artificial Intelligence, and Machine & Deep Learning. Experience a unique opportunity to test out the latest performance optimized frameworks and tools, advanced coding knowledge and best practices to get started implementing AI guided by experts from Intel®.
Target Audience:
Data Scientist, application developers and HPC benchmarkers targeting the deep learning and machine learning domain.
Prerequisites:
Beginning to intermediate level of domain AI knowledge.
Basic skills of programming , ideally some Python knowldege
Type of tutorial: Mix of lectures and hands-on tutorials
Special Requirements:
Attendees should bring their laptop with an SSH- & VNC client
Attendees will get for hands on-labs also access to the CHPC cluster
Outline of full syllabus:
08:00 Registration
09:00 Introduction
• Introduction round & Agenda
• Introduction Intel Software Developer Tools
• Introduction to Machine Learning / Deep Learning
10:30 Morning Refreshment Break
11:00
Classic Machine Learning Tools
• Intel performance Libraries - MKL & DAAL
• Intel Distribution for Python (IDP) Introduction
• IDP Hands on labs
o NumPy & MKL
o K-Means Clustering & DAAL
o SVM & DAAL
12:30 Lunch
13:30
Deep Learning (DL) Tools
• Intel performance Libraries for DL – MKL-DNN & MLSL
• Intel optimized Frameworks / TensorFlow
o TensorFlow Image Classification Hands-on Lab
o Introduction simple CNN
o Monitored Training Session
15:00 Afternoon Refreshment Break
15:30
Deep Learning Tools (cont’d)
• Intel optimized Frameworks / TensorFlow (cnt’d)
o Horovod – distributed classification
o Importing external Images
o Custom batches
• Benchmarking distributed TensorFlow (BKMs)
• Deep Learning Scaling – large scale results (BigDL/Spark)
• Wrap-Up & Q&A
17:00 End of Day
A dedicated Chemistry and Material Science SIG session consisting of keynotes, ordinary oral presentations, student oral presentations as well as flash poster presentations. The program will also include the annual SIG meeting.
Program:
09:00 Keynote: John Mack, Rhodes University, The use of Gaussian 09 to identify trends in the optical properties and electronic structures of porphyrin and BODIPY dyes
09:30 Talk: Marianne Conradie, UFS, Determination of the steric effect of a bulky R substituent on the reaction mechanism of [Rh(CH3COCRCOCH3)(CO)(PPh3)] + CH3I
09:50 Talk: Cliffton Masedi, UL, Computer Simulation and Phase Diagram Prediction of Li2S1-xSex Systems
10:10 Talk: Thishana Singh, DUT, Biosynthesis and computational analysis of amine-ended dual thiol ligand functionalized gold nanoparticles
10:30 Morning Tea Break
11:00 Student talk: Cameron Matthews, NMU, A theoretical study of electronic (QTAIM NBO) and geometric structures of a series of lanthanide trichloride LnCl3 complexes (Ln=La-Lu) with the N-donor ligand bis(2-pyridylmethyl)amine (DPA)
11:15 Student talk: Elkana Rugut, Wits, Thermoelectric properties of CdAl2O4 spinel
11:30 Student talk: Fisayo Olotu, UKZN, Delving through the wonders of green tea - Probing the dynamic selectivity of polyphenol Epigallocatechin gallate
11:45 Student talk: Francis Opoku, UJ, Tuning the electronic structures
12:00 Student poster flash presentation: Ramogohlo Diale, UL, Computational modelling of TiPd-Ir high temperature shape memory alloys
12:05 Student poster flash presentation: Sharlene-Asia Naicker, UKZN, Computational studies of the corrosion mechanisms in transformers
12:10 Student poster flash presentation: Surya Narayana Maddila, UKZN, Synthesis and spectral properties of 5-substituted 1H-tetrazole derivatives using with Mn/HAp catalyst and DFT calculations
12:15 Student poster flash presentation: Nkosinathi Malaza, Wits, Double-layer capacitance from ab initio simulations
12:20 Student poster flash presentation: Jonathan Gertzen, UCT, MAX Phases as an Electrocatalyst Support Material: A DFT Study
12:25 Discussion
12:30 Lunch Break
13:30 Talk: Nkululeko Damoyi, MUT, A DFT study of the ODH of n-hexane over isolated H3VO4 and H4V2O7
13:50 Student talk: Adebayo Adeniyi, UFS, The theoretical investigation of reduction potentials and and spectroscopic properties of nitrobenzene and keto-enol molecules
14:05 Student talk: Abdulgaffar, UP, Hybrid functional study of BiOi, BiCs, BiBsHi and BiOiOi Complexes in Silicon
14:20 Student talk: Joseph Simfukwe, UP, First Principles study of Zn/Cu doped hematite surfaces for Photoelectrochemical water splitting
14:35 Talk: Khalid Ahmed, UKZN, Understanding the influence of pH/pKa by theoretical and computational method for loading/release of Ibuprofen (To Be Confirmed)
14:55 Discussion
15:00 Afternoon Tea Break
15:30 SIG Meeting
HPC Carpentry is an introduction to working on the CHPC cluster with an overview of the concepts necessary to make effective use of a high performance computing (HPC) cluster along with practical introduction to submitting jobs, finding software and splitting up your data so that it can be processed in parallel. This workshop is appropriate for people who are new to the world of clusters and want a practical introduction to the benefits they can derive from HPC and high throughput computing (HTC).
Molecular Dynamics and analysis can be complicated for novitiates and researchers from neighbouring disciplines. Building upon the Galaxy Project platform, BRIDGE (Biomolecular Reaction & Interaction Dynamics Global Environment) is a web application that provides the ability to get started running molecular dynamics and analyses using curated workflows.
Target Audience: Computational chemists, biologists, protein scientists, crystallographers and anyone interested in running and analysing molecular dynamics simulations who is not familiar with the command line.
Prerequisites: Basic understanding of chemistry. A basic understanding of or keen interest in molecular dynamics.
Special requirements: Bring a laptop. Have access to the CHPC cluster. Register with BRIDGE. Install molecular viewer e.g. VMD.
Day: Thursday (6 Dec)
Duration: 1 day
Size: 15 seats
Description:
Join us for one day of hands-on sessions on Artificial Intelligence, and Machine & Deep Learning. Experience a unique opportunity to test out the latest performance optimized frameworks and tools, advanced coding knowledge and best practices to get started implementing AI guided by experts from Intel®.
Target Audience:
Data Scientist, application developers and HPC benchmarkers targeting the deep learning and machine learning domain.
Prerequisites:
Beginning to intermediate level of domain AI knowledge.
Basic skills of programming , ideally some Python knowldege
Type of tutorial: Mix of lectures and hands-on tutorials
Special Requirements:
Attendees should bring their laptop with an SSH- & VNC client
Attendees will get for hands on-labs also access to the CHPC cluster
Outline of full syllabus:
08:00 Registration
09:00 Introduction
• Introduction round & Agenda
• Introduction Intel Software Developer Tools
• Introduction to Machine Learning / Deep Learning
10:30 Morning Refreshment Break
11:00
Classic Machine Learning Tools
• Intel performance Libraries - MKL & DAAL
• Intel Distribution for Python (IDP) Introduction
• IDP Hands on labs
o NumPy & MKL
o K-Means Clustering & DAAL
o SVM & DAAL
12:30 Lunch
13:30
Deep Learning (DL) Tools
• Intel performance Libraries for DL – MKL-DNN & MLSL
• Intel optimized Frameworks / TensorFlow
o TensorFlow Image Classification Hands-on Lab
o Introduction simple CNN
o Monitored Training Session
15:00 Afternoon Refreshment Break
15:30
Deep Learning Tools (cont’d)
• Intel optimized Frameworks / TensorFlow (cnt’d)
o Horovod – distributed classification
o Importing external Images
o Custom batches
• Benchmarking distributed TensorFlow (BKMs)
• Deep Learning Scaling – large scale results (BigDL/Spark)
• Wrap-Up & Q&A
17:00 End of Day
Description:
Join us for one day of hands-on sessions on Artificial Intelligence, and Machine & Deep Learning. Experience a unique opportunity to test out the latest performance optimized frameworks and tools, advanced coding knowledge and best practices to get started implementing AI guided by experts from Intel®.
Target Audience:
Data Scientist, application developers and HPC benchmarkers targeting the deep learning and machine learning domain.
Prerequisites:
Beginning to intermediate level of domain AI knowledge.
Basic skills of programming , ideally some Python knowldege
Type of tutorial: Mix of lectures and hands-on tutorials
Special Requirements:
Attendees should bring their laptop with an SSH- & VNC client
Attendees will get for hands on-labs also access to the CHPC cluster
Outline of full syllabus:
08:00 Registration
09:00 Introduction
• Introduction round & Agenda
• Introduction Intel Software Developer Tools
• Introduction to Machine Learning / Deep Learning
10:30 Morning Refreshment Break
11:00
Classic Machine Learning Tools
• Intel performance Libraries - MKL & DAAL
• Intel Distribution for Python (IDP) Introduction
• IDP Hands on labs
o NumPy & MKL
o K-Means Clustering & DAAL
o SVM & DAAL
12:30 Lunch
13:30
Deep Learning (DL) Tools
• Intel performance Libraries for DL – MKL-DNN & MLSL
• Intel optimized Frameworks / TensorFlow
o TensorFlow Image Classification Hands-on Lab
o Introduction simple CNN
o Monitored Training Session
15:00 Afternoon Refreshment Break
15:30
Deep Learning Tools (cont’d)
• Intel optimized Frameworks / TensorFlow (cnt’d)
o Horovod – distributed classification
o Importing external Images
o Custom batches
• Benchmarking distributed TensorFlow (BKMs)
• Deep Learning Scaling – large scale results (BigDL/Spark)
• Wrap-Up & Q&A
17:00 End of Day