Tuning parallel applications on software-defined supercomputers. Supercomputers are used by many Australian industries and laboratories to make better products and perform critical predictions, and it is essential that codes operate efficiently. This project aims to assist programmers in identifying performance bottlenecks in their code quickly and easily. The project expects to supersede the current methods, which are often complex and time-consuming, by developing innovative software tools and ....Tuning parallel applications on software-defined supercomputers. Supercomputers are used by many Australian industries and laboratories to make better products and perform critical predictions, and it is essential that codes operate efficiently. This project aims to assist programmers in identifying performance bottlenecks in their code quickly and easily. The project expects to supersede the current methods, which are often complex and time-consuming, by developing innovative software tools and techniques. The expected outcomes include novel software, verified by industry partners in real world case studies, ranging from life sciences to hypersonic transport. This should provide significant benefits, including the capacity for Australian industries to access world-class supercomputing technology.Read moreRead less
Finding concurrency bugs in multithreaded software. This project aims to develop sound and practical techniques for detecting and eliminating concurrency bugs for object-oriented languages like Java, enabled by a new model for concurrent effects. The expected outcome is a novel technology that will significantly improve the safety, productivity and efficiency of large-scale concurrent programming.
Software debuggers for next generation heterogeneous supercomputers. Supercomputing underpins a wide range of areas of importance to the Australian economy; mining, agriculture, engineering and medical research to name a few. It is of critical importance that software solutions in these areas behave correctly. This project will develop software tools and techniques to help locate errors in such applications.
Symbolic synthesis of knowledge-based program implementations. Systems with concurrent streams of activity are ubiquitous in computer hardware and software designs, but are conceptually complex, and fraught with faults and inefficiency. The project aims to address these difficulties by automating aspects of system design, to relieve the designer of the need to reason about complex patterns of information flow.
Verified concurrent memory management on modern processors. This project aims to formally verify automatic memory managers in the presence of concurrency and the weakly ordered memory of modern processors. A new framework for verifying memory managers, reusable for a wide range of managed programming languages, target hardware, policies, and algorithms will be developed. Expected technical outcomes include improved techniques to ensure trustworthiness of the foundations on which critical softwar ....Verified concurrent memory management on modern processors. This project aims to formally verify automatic memory managers in the presence of concurrency and the weakly ordered memory of modern processors. A new framework for verifying memory managers, reusable for a wide range of managed programming languages, target hardware, policies, and algorithms will be developed. Expected technical outcomes include improved techniques to ensure trustworthiness of the foundations on which critical software infrastructures are built. This will significantly enhance the security of public and private cyber assets, and deliver applications that are more robust and trustworthy, across a range of critical infrastructure such as transportation, communication, energy and defence.Read moreRead less
Automatic Energy Tuning of Parallel Applications on a Hybrid Supercomputer. Energy efficiency is a critical challenge in building next-generation supercomputers. This project aims to provide a new energy-tuning tool integrated with Cray’s systems, to simplify the process of tuning hybrid applications and managing efficient energy use. Although hardware components play a dominant role in saving energy, heterogeneous systems offer the opportunity to exploit the extremely high concurrency with mode ....Automatic Energy Tuning of Parallel Applications on a Hybrid Supercomputer. Energy efficiency is a critical challenge in building next-generation supercomputers. This project aims to provide a new energy-tuning tool integrated with Cray’s systems, to simplify the process of tuning hybrid applications and managing efficient energy use. Although hardware components play a dominant role in saving energy, heterogeneous systems offer the opportunity to exploit the extremely high concurrency with modest energy consumption using accelerators. Accordingly, the future of parallel computing must consider the trade-off between obtaining the optimal performance and the allowed power budget. The project plans to design parallel programming environments that support energy analysis and tuning.Read moreRead less
Edge-Accelerated Deep Learning. Implementing deep learning (DL) applications usually requires a large amount of collected data and powerful computing resources in the cloud. However, this centralised approach has issues of high latency, large bandwidth usage, and possible privacy violation for many practical applications. Without properly addressing these issues, the wider application of DL in practice will seriously be hindered. This project aims to solve several key challenging problems in eff ....Edge-Accelerated Deep Learning. Implementing deep learning (DL) applications usually requires a large amount of collected data and powerful computing resources in the cloud. However, this centralised approach has issues of high latency, large bandwidth usage, and possible privacy violation for many practical applications. Without properly addressing these issues, the wider application of DL in practice will seriously be hindered. This project aims to solve several key challenging problems in effective deployment and efficient execution of DL applications in a distributed edge-computing environment. Several innovative edge-computing methods will be developed for DL training, inference and implementation to achieve high performance with low latency and enhanced privacy.Read moreRead less