Heterogeneous Thinking
Lead Research Organisation:
University of Edinburgh
Department Name: Sch of Informatics
Abstract
In 1965 Gordon Moore observed that the number of components on a computer chip doubles every 18 months. For almost 50 years we have enjoyed this exponentially increasing computer power. This has transformed society, heralding today's computer-age.
This growth is based on a fundamental contract between hardware and software that until recently has rarely been questioned. The contract is: hardware may change radically "under the hood", but the code you ran on yesterday's machine will run just the same on tomorrow's - but even faster. Hardware may change, but it looks the same to software, always speaking the same language. This common, consistent language allows the decoupling of software development from hardware development. It has allowed programmers to invest significant effort in software development, secure in the knowledge that it will have decades of use. Alarmingly, this contract will begin to fall apart,
putting in jeopardy the massive investment in software.
The reason for the breakdown is due to the end of Moore's Law. Technology can no longer be relied upon to scale smaller and provide performance; the end of Moore's Law is beginning to force new approaches to computer design. The cost of maintaining the common consistent language contract is enormous. If we break the contract and develop specialised hardware, there is potentially a 10,000 x massive performance gain available. For this reason, it is clear that
in future, hardware will be increasingly specialised and heterogeneous. Currently, however, there is no clear way of programming and using such hardware.
As it stands, either hardware evolution will stall as software cannot fit or software will be unable to exploit hardware innovation. Such a crisis requires a fundamental re-think of how we design, program and use heterogeneous systems.
What we need is an approach that liberates hardware from the uniform language contract and efficiently connects existing and future software to the emerging heterogeneous landscape.
This project proposes a way of doing this by re-thinking how we connect software and hardware by a more flexible language interface which can change from one processor to the next. It allows existing software to use future hardware and allows hardware innovation to connect to new and emerging applications areas such robotics, augmented reality and deep learning. If successful, it will usher in an era of change in systems design where, rather than deny and fear the end of
Moore's law, we embrace and exploit it.
This growth is based on a fundamental contract between hardware and software that until recently has rarely been questioned. The contract is: hardware may change radically "under the hood", but the code you ran on yesterday's machine will run just the same on tomorrow's - but even faster. Hardware may change, but it looks the same to software, always speaking the same language. This common, consistent language allows the decoupling of software development from hardware development. It has allowed programmers to invest significant effort in software development, secure in the knowledge that it will have decades of use. Alarmingly, this contract will begin to fall apart,
putting in jeopardy the massive investment in software.
The reason for the breakdown is due to the end of Moore's Law. Technology can no longer be relied upon to scale smaller and provide performance; the end of Moore's Law is beginning to force new approaches to computer design. The cost of maintaining the common consistent language contract is enormous. If we break the contract and develop specialised hardware, there is potentially a 10,000 x massive performance gain available. For this reason, it is clear that
in future, hardware will be increasingly specialised and heterogeneous. Currently, however, there is no clear way of programming and using such hardware.
As it stands, either hardware evolution will stall as software cannot fit or software will be unable to exploit hardware innovation. Such a crisis requires a fundamental re-think of how we design, program and use heterogeneous systems.
What we need is an approach that liberates hardware from the uniform language contract and efficiently connects existing and future software to the emerging heterogeneous landscape.
This project proposes a way of doing this by re-thinking how we connect software and hardware by a more flexible language interface which can change from one processor to the next. It allows existing software to use future hardware and allows hardware innovation to connect to new and emerging applications areas such robotics, augmented reality and deep learning. If successful, it will usher in an era of change in systems design where, rather than deny and fear the end of
Moore's law, we embrace and exploit it.
Planned Impact
Solving the heterogeneous crisis will clearly have a major impact on society. It will safeguard and future-proof the massive $400B investment in today's business software systems and provide a pathway to greater performance in the future. It provides a roadmap and a solution to managing the decline of Moore's Law, enabling innovation and continued growth in computer systems.
Its immediate effect will be first realised in the academic world, described above followed by industry and then society more broadly.
Industrial Impact
Clearly, companies that build compiler tools for the programming of parallel and heterogeneous platform will be able to exploit the research developed in this project. The automated discovery of parallelism is an immediate starting point for technology transfer. Such research can be exploited by companies such as our partners CodePlay and ARM or others such as CriticalBlue, to provide expert programmer advice and assistance.
The ability to evaluate the market impact of a new processor IP design will be of enormous benefit to IP designers such as Imagination Technology, Cirrus Logic and our partner ARM. This project will allow them to determine how well a proposed design fits a target application domain. The ability to include ease of use in the design process will be of real significant benefit to IP designers.
In the longer term, this project will impact application developers, especially those engaged in vision based systems, robotics and AI more generally. Companies investing in AI based applications such as Microsoft, Disney, Toshiba, Sony and Dyson will be able to explore the impact of new hardware on their products and use my research to guide the design of their software.
Societal Impact
Technological change and disruption has a profound impact on society, though its impact is felt much later. As this project is a fundamental computer science project, its impact, although wide, is often indirect.
Ultimately, this project, if successful, ensures that improved computer performance will be possible despite the seismic changes due to the end of Moore's Law. This means the investment and jobs associated with the world's software base is more assured. The digital fabric that we increasingly rely upon will continue to grow and improve rather than falter and atrophy. There will be increased economic benefits with existing system providers continuing to thrive and grow. Due to new opportunities provided by an increase in performance, new companies will be created especially around emerging AI based applications.
As well as tangible economic benefits, society will also be impacted indirectly by the enabling action computers have on important aspects of our daily lives. For example, by increasing computer power this project impacts medicine, where it can help areas as diverse as medical robotics to longitudinal patient record analysis. In the completely different area of entertainment, such increased performance allows immersive realtime augmented reality. A major challenge for fundamental computer science is communicating such impact to a wider audience. An important part of the project is to devise ways to engage with the public and seek a dialogue on the impact of technological change.
To deliver this impact we will do the following:
- Publish Research in the most respected journals and conferences
- Distribute demonstrator systems to academics and industrial partners
- Organise international workshops to disseminate and discuss findings
- Engage in bilateral academic exchange visits
- Hold a high profile industrial innovation workshop
- Transfer technology to industrial partners.
- Explore spin-out company formation
- Communicate with the public facilitated by the University's public engagement team
- Promote achievements to general science audiences at the Edinburgh Science Festival and Beltane events
Its immediate effect will be first realised in the academic world, described above followed by industry and then society more broadly.
Industrial Impact
Clearly, companies that build compiler tools for the programming of parallel and heterogeneous platform will be able to exploit the research developed in this project. The automated discovery of parallelism is an immediate starting point for technology transfer. Such research can be exploited by companies such as our partners CodePlay and ARM or others such as CriticalBlue, to provide expert programmer advice and assistance.
The ability to evaluate the market impact of a new processor IP design will be of enormous benefit to IP designers such as Imagination Technology, Cirrus Logic and our partner ARM. This project will allow them to determine how well a proposed design fits a target application domain. The ability to include ease of use in the design process will be of real significant benefit to IP designers.
In the longer term, this project will impact application developers, especially those engaged in vision based systems, robotics and AI more generally. Companies investing in AI based applications such as Microsoft, Disney, Toshiba, Sony and Dyson will be able to explore the impact of new hardware on their products and use my research to guide the design of their software.
Societal Impact
Technological change and disruption has a profound impact on society, though its impact is felt much later. As this project is a fundamental computer science project, its impact, although wide, is often indirect.
Ultimately, this project, if successful, ensures that improved computer performance will be possible despite the seismic changes due to the end of Moore's Law. This means the investment and jobs associated with the world's software base is more assured. The digital fabric that we increasingly rely upon will continue to grow and improve rather than falter and atrophy. There will be increased economic benefits with existing system providers continuing to thrive and grow. Due to new opportunities provided by an increase in performance, new companies will be created especially around emerging AI based applications.
As well as tangible economic benefits, society will also be impacted indirectly by the enabling action computers have on important aspects of our daily lives. For example, by increasing computer power this project impacts medicine, where it can help areas as diverse as medical robotics to longitudinal patient record analysis. In the completely different area of entertainment, such increased performance allows immersive realtime augmented reality. A major challenge for fundamental computer science is communicating such impact to a wider audience. An important part of the project is to devise ways to engage with the public and seek a dialogue on the impact of technological change.
To deliver this impact we will do the following:
- Publish Research in the most respected journals and conferences
- Distribute demonstrator systems to academics and industrial partners
- Organise international workshops to disseminate and discuss findings
- Engage in bilateral academic exchange visits
- Hold a high profile industrial innovation workshop
- Transfer technology to industrial partners.
- Explore spin-out company formation
- Communicate with the public facilitated by the University's public engagement team
- Promote achievements to general science audiences at the Edinburgh Science Festival and Beltane events
Publications
Wang Z
(2018)
Machine Learning in Compiler Optimization
in Proceedings of the IEEE
Patacchiola M
(2020)
Bayesian Meta-Learning for the Few-Shot Setting via Deep Kernels
Collie B
(2020)
M 3
Ginsbach P
(2020)
Automatically harnessing sparse acceleration
Kaszyk K
(2019)
Full-System Simulation of Mobile CPU/GPU Platforms
Crawford L
(2019)
Specialization Opportunities in Graphical Workloads
Collie B
(2021)
Program Lifting using Gray-Box Behavior
Description | The ability to understand hardware behaviour without access to any knowledge of its design or internals. This gives us the potential to port software to any new hardware device |
Exploitation Route | Developing new hardware designs |
Sectors | Digital/Communication/Information Technologies (including Software) Electronics |
Description | Raising the abstraction level of code so that it can be run more efficiently on new hardware |
First Year Of Impact | 2023 |
Sector | Digital/Communication/Information Technologies (including Software) |
Impact Types | Economic |
Description | Software defined hardware |
Amount | £9,000,000 (GBP) |
Funding ID | Transmuter |
Organisation | Defense Advanced Research Projects Agency (DARPA) |
Sector | Public |
Country | United States |
Start | 09/2018 |
End | 04/2021 |
Description | ARM Research Centre of Excellence in Heterogeneous Computing 2019 |
Organisation | Arm Limited |
Country | United Kingdom |
Sector | Private |
PI Contribution | Heterogeneous Computing |
Collaborator Contribution | Next generation GPU IP at RTL and ISA level |
Impact | 5 new PhD studentd |
Start Year | 2019 |
Description | Huawei Lab Edinburgh |
Organisation | Huawei Technologies Research and Development UK Ltd |
Country | United Kingdom |
Sector | Private |
PI Contribution | PhD Supervision |
Collaborator Contribution | Money |
Impact | None so far |
Start Year | 2021 |
Description | Key note at ACM/IEEE CGO Washington Feb 2019 |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Industry/Business |
Results and Impact | Presenting the vision of my work to a mixed academic/business/government agency audience |
Year(s) Of Engagement Activity | 2019 |
URL | http://cgo.org/cgo2019/ |
Description | Microsoft Research Talk Seatle |
Form Of Engagement Activity | A formal working group, expert panel or dialogue |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Industry/Business |
Results and Impact | Position talk on heterogeneity |
Year(s) Of Engagement Activity | 2019 |