Skip to main content

Coping in a parallel world

The heat produced by modern CPUs means that chip designers can no longer crank up the clock speed. But they can fit more processing cores on a chip. 2-core and 4-core x86 chips are already standard; Sun's SPARC chips have 8 or 16 cores, and Intel recently announced an 80-core prototype.

To get the best performance out of multi-core chips, especially once you get beyond 2 or 4 cores, you either have to write applications to make best use of those cores or write compilers and dynamic optimising systems that automatically transform your code to get that performance. The problem is, most programmers aren't very good at this and most of the tools they use aren't brilliant either.

I think this has immediate policy implications. CS departments should immediately, if they haven't already done so, make the teaching of parallel and distributed programming a required component of their undergraduate courses. Companies and professional bodies should encourage their staff to retrain. For this retraining, someone has to prepare and give the courses. Some government funding might kickstart the process and so ensure that the UK doesn't fall behind.

Several pundits, including Tony Hey of Microsoft, have suggested that experience from the supercomputer world could help, because supercomputers have been massively parallel for years. There is some truth in this, and organisations such as EPCC are already teaching parallel programming techniques. But there are differences too. The economics of the supercomputer world are rather old fashioned; there, the computer is still the expensive part of the system while programmers' time is relatively cheap. So programmers often tune each application for each new machine.

You can see this philosophy in much of the scientific grid world: people submit jobs that require a certain number of processors to run. Think about that. They're not asking for a certain amount of processing power, nor for their jobs to run in a certain time or for a certain cost. Their applications are programmed to run on a specific number of processors. To anyone outside that community, this approach is clearly crazy, and it certainly won't transfer to a world where multi-core processors are plentiful and programmers' time is expensive.

So the distributed (grid) world and the multicore CPU world face similar problems. Both need new programming models and tools. Fortunately there are many computer scientists who are investigating better ways to program distributed and parallel machines. I can't attempt to represent the field but some of my friends and colleagues are working on such problems. For example, Murray Cole at the University of Edinburgh has developed a model for transforming parallel programs on to different numbers of processors. In 2002, a team from Microsoft Research in Cambridge incorporated modern concurrency abstractions into a research version of C# - replacing the usual locks, semaphores and critical regions that Tony Hoare invented 40 year ago.

Here at the e-Science Institute, we are about to launch a theme - a series of workshops and visitors - on Distributed Programming Abstractions. I certainly hope that this will address some of the questions raised above and that soome of the people reading this will contribute. Another useful resource is this wiki at Berkeley, which also includes an interesting white paper.

Comments

I wholeheartedly agree with the challenge that Dave offers here. We are building a largescale, distributed computing infrastructure capable of wholesale parallel operation for both transactional and computational computing. It should be a matter of course that significant applications are designed with multiple-thread operation in mind. This is not a new problem in CS, its just that we've never had much opportunity in the past to do this for real!!
Dave Berry said…
Intel have a white paper than makes these points and many more, at ftp://download.intel.com/research/platform/terascale/terascale_overview_paper.pdf.

Thoams Sterling addresses these issues too. His recent keynote points out some of the differences as well as the similarities. A major difference is the bottleneck in memory access on multi-core CPUs. See http://www.cct.lsu.edu/~tron/ICCC06EndNoteFinal.pdf.

Popular posts from this blog

Webinar: Powering your business with Cloud Computing

On October 14th, I will be hosting a Grid Computing Now! web seminar on the topic of Cloud Computing. We have lined up two very interesting speakers who are using Cloud now to make businesses work. Ross Cooney had a good technological solution to sell but couldn't make it economic until Cloud Computing allowed him to pay for his computation only when he needed it. He will discuss the instant benefits and long term impact of cloud computing to the development, competitiveness and scalability of your application. Alan Williamson created the BlueDragon Java CFML runtime engine that powers MySpace.com. He advises several businesses and will give an overview of the different types of services available and how to avoid being locked-in to a single supplier. You can register for this event here .

Business Model Canvas

A Business Model Canvas is a tool for mapping the core functions and capabilities of an organisation.  Compared to the Core Diagrams that I described in an earlier post , the business model canvas attempts to present more aspects of the business, starting with the value proposition – a statement of what the organisation offers to its users (in the business world, to its customers).  It shows the activities and resources, as Core Diagrams do, but also shows user relationships & channels, and also benefits and costs.  I’m not aware of any universities that have used this tool but you can find examples from elsewhere on the web. We are considering business model canvases as a tool for mapping the strategic capabilities of units at the University of Edinburgh.  Phil Taylor, our EA contractor, sketched an outline of what a business model canvas might begin to look like for HR: This is only intended to be suggestive: the real canvas would need to result from in-depth discussions abo

Changing Principles

In EA, architecture principles set a framework for making architectural decisions.  They help to establish a common understanding across different groups of stakeholders, and provide guidance for portfolios and projects.  Michael Durso of the LSE gave a good introduction to the idea in a webinar last week for the UCISA EA community. Many organisations take the TOGAF architecture principles as a starting point.  These are based on the four architectural domains of TOGAF: business, information/data, applications, technology/infrastructure.  These principles tend to describe what should be done, e.g. re-use applications, buy in software rather than build it, keep data secure.  See for example the principles adopted at Plymouth University and the University of Birmingham . Recently though, I encountered a different way of looking at principles.  The user experience design community tend to focus more on how we should do things.  E.g. we should start with user needs, use iterat