A Programming Model of Object Validity for Secure and Efficient Concurrency. To provide improved performance and security for software applications, Australia's ICT industry must adapt to the complex programming demands of modern multicore processors. The programming model developed in the project represents a breakthrough solution. A seamless integration of concurrency and object orientation leads to a simple yet powerful programming style that is compatible with today's premier approach to bui ....A Programming Model of Object Validity for Secure and Efficient Concurrency. To provide improved performance and security for software applications, Australia's ICT industry must adapt to the complex programming demands of modern multicore processors. The programming model developed in the project represents a breakthrough solution. A seamless integration of concurrency and object orientation leads to a simple yet powerful programming style that is compatible with today's premier approach to building large-scale software systems with significantly improved performance and security. The product will reduce development time for compute-intensive applications in many industry sectors, including health care (e.g. patient monitoring), finance, defence, environment, mining, manufacturing and computer games.Read moreRead less
Analysis and Optimisation of Incomplete Object-Oriented Programs. This project will push Australia to the forefront of OO technologies and will increase the profile of advanced programming language design and implementation research in Australia. The completed framework will
likely be commercially viable as Australia industries such as banks and
insurance companies and government organisations hosting large
OO software systems may benefit (in terms of performance and
code security) from the ....Analysis and Optimisation of Incomplete Object-Oriented Programs. This project will push Australia to the forefront of OO technologies and will increase the profile of advanced programming language design and implementation research in Australia. The completed framework will
likely be commercially viable as Australia industries such as banks and
insurance companies and government organisations hosting large
OO software systems may benefit (in terms of performance and
code security) from the technology developed as a result of this research. The completed framework will also provide an excellent platform for researchers in Australia and elsewhere to do research in OO technologies, helping in producing quality students for the Australian IT industry.Read moreRead less
Scratchpad-based Memory Allocation Techniques for Embedded Software. This research aims at developing automatic memory allocation algorithms to maximise the effective utilisation of scratchpad memories, which will lead to significantly improved performance and energy usage in embedded applications. The outcomes of this project will provide generic solutions to many Australia-based industries, including telecommunication, network management, sensor networks, automotive and instrumentation/measure ....Scratchpad-based Memory Allocation Techniques for Embedded Software. This research aims at developing automatic memory allocation algorithms to maximise the effective utilisation of scratchpad memories, which will lead to significantly improved performance and energy usage in embedded applications. The outcomes of this project will provide generic solutions to many Australia-based industries, including telecommunication, network management, sensor networks, automotive and instrumentation/measurement, where embedded systems are ubiquitously used. Therefore, this project will significantly contribute to the Priority Area (Frontier Technologies for Building and Transforming Australian Industries), hence fits into its Priority Goal: Frontier Technologies.Read moreRead less
Compiler-Directed Code Tiling for Higher Program Performance and Predictability on Multi-Level Memory Hierarchies. Programming languages invariably induce a perception of memory as being flat, whereas actual computers are being equipped increasingly with deep memory hierarchies to overcome the ever-widening performance gap between processors and memories. This mismatch can result in low locality of reference and poor performance. This project will develop a new array layout transformation, calle ....Compiler-Directed Code Tiling for Higher Program Performance and Predictability on Multi-Level Memory Hierarchies. Programming languages invariably induce a perception of memory as being flat, whereas actual computers are being equipped increasingly with deep memory hierarchies to overcome the ever-widening performance gap between processors and memories. This mismatch can result in low locality of reference and poor performance. This project will develop a new array layout transformation, called data tiling, used in tandem with loop tiling to achieve significantly improved performance gains (and predictability) for array-dominated applications on multi-level memory hierarchies. We will explore this joint restructuring to obtain higher performance and predictability for whole programs, kernel libraries, iterative stencil codes and embedded applications.
Read moreRead less
A safe and efficient multi-language component framework based on dynamic compilation. Components are a versatile mechanism for the construction of extendible systems and for improved code reusability. Emerging environments like Microsoft's .NET support the interaction of components written in different programming languages. To date, such systems do not adequately isolate components which therefore must trust each other. This is unacceptable for components loaded across the Internet. We will dev ....A safe and efficient multi-language component framework based on dynamic compilation. Components are a versatile mechanism for the construction of extendible systems and for improved code reusability. Emerging environments like Microsoft's .NET support the interaction of components written in different programming languages. To date, such systems do not adequately isolate components which therefore must trust each other. This is unacceptable for components loaded across the Internet. We will develop a framework for secure execution of component software, based on hardware-enforced protection domains. To minimize the resulting runtime overheads we will provide a component model with lightweight context switching. A novel typed intermediate language and dynamic compilation techniques will offer complementary optimisations.
Read moreRead less
Micro Compilers: An Extensible Compiler Architecture for Increased Flexibility and Safety. Programming languages specialised to a particular application domain can cut software development costs and reduce programming errors. Unfortunately, most application domains cannot sustain the costly development and maintenance of the specialised compilers required to implement specialised languages. We address this problem by introducing a novel customisable compiler architecture that can be adapted to ....Micro Compilers: An Extensible Compiler Architecture for Increased Flexibility and Safety. Programming languages specialised to a particular application domain can cut software development costs and reduce programming errors. Unfortunately, most application domains cannot sustain the costly development and maintenance of the specialised compilers required to implement specialised languages. We address this problem by introducing a novel customisable compiler architecture that can be adapted to specialised languages and other special-purpose compiler requirements. Customisable compilers are especially important to a country like Australia that has only limited resources for special purpose developments. Moreover, we will train students at undergraduate and postgraduate level in the area of programming languages and compilers.Read moreRead less
Portable High-Performance Computing Based on Flattening and Fusion. We aim at simplifying the development and improving the portability of computing intensive applications, such as those from computational science and engineering, which are of growing scientific and economic relevance (e.g., drug synthesis, virtual wind tunnel, and geologic surveys). Our focus is on achieving portable high-performance for a particularly expressive high-level notation supporting irregular algorithms (like sparse ....Portable High-Performance Computing Based on Flattening and Fusion. We aim at simplifying the development and improving the portability of computing intensive applications, such as those from computational science and engineering, which are of growing scientific and economic relevance (e.g., drug synthesis, virtual wind tunnel, and geologic surveys). Our focus is on achieving portable high-performance for a particularly expressive high-level notation supporting irregular algorithms (like sparse matrices and hierarchical N-body codes). We will develop a set of program transformations and integrate them into a compiler implementing an aggressive fusion strategy optimising for the memory hierarchy. We do not propose a new programming language, but work within an existing one.Read moreRead less
Dataparallel Programming for Multicore Processors. The proposed project will contribute to the development of frontier technologies to help build Australian industries. The project is designed to unlock significant performance improvements with current and future multicore (processor) computer architectures. This potential performance improvement can be achieved with parallel programming models. This is crucial for ICT applications in performance hungry areas, such as biotechnology, finance, mul ....Dataparallel Programming for Multicore Processors. The proposed project will contribute to the development of frontier technologies to help build Australian industries. The project is designed to unlock significant performance improvements with current and future multicore (processor) computer architectures. This potential performance improvement can be achieved with parallel programming models. This is crucial for ICT applications in performance hungry areas, such as biotechnology, finance, multi-media, and 'info-tainment'. The project will also provide research training and increase local expertise in parallel programming for multicore processors, an area that is quickly growing in importance.Read moreRead less
Towards High-performance and Fault-tolerant Distributed Java Implementations. Java Virtual Machines form an important part of the web and business
server market. Distributed Java Virtual Machines have the potential to
make a significant contribution to industries that utilize this
technology. An attractive platform for this purpose is the
cluster, a highly cost-effective and scalable parallel computer
model. However, realizing on such a platform a high performance virtual
machine implem ....Towards High-performance and Fault-tolerant Distributed Java Implementations. Java Virtual Machines form an important part of the web and business
server market. Distributed Java Virtual Machines have the potential to
make a significant contribution to industries that utilize this
technology. An attractive platform for this purpose is the
cluster, a highly cost-effective and scalable parallel computer
model. However, realizing on such a platform a high performance virtual
machine implementation tolerant to hardware and software faults, and
having efficient memory utilization, presents many challenging research
issues. This project will address these issues by extending a highly
efficient and extensible Java implementation to be aware of its cluster
environment.
Read moreRead less
Extending a family of garbage collectors. Garbage collection is a key component in the automatic management of storage in computer systems. It is an essential property of modern programming systems that frees the programmer from a significant error-prone task. Our interest is in garbage collection in distributed systems involving a number of networked computers. Using our novel construction methodology, we have jointly produced a family of collection algorithms that are significantly simpler and ....Extending a family of garbage collectors. Garbage collection is a key component in the automatic management of storage in computer systems. It is an essential property of modern programming systems that frees the programmer from a significant error-prone task. Our interest is in garbage collection in distributed systems involving a number of networked computers. Using our novel construction methodology, we have jointly produced a family of collection algorithms that are significantly simpler and more efficient than previous work. Here we wish to extend this family to operate effectively in a specific architecture increasingly favoured by many modern distributed high-performance computing systems.Read moreRead less