The Next Big Thing
Posts about next generation technologies and their effect on business.

Multi-core, where will it end?

A number of times I've mentioned the rise of parallel processing and multi-core as well as the effect on development and the kinds of solutions that are consumed. I was talking with some folks at TI the other day about the projections of multi-core processors on the market and the attached chart was discussed.


If you place this information on a logarithmic scale it definitely appears to be exponential in nature and we all know examples of the shift in thinking that exponential growth causes - essentially a shift to a view to one of abundance and not scarcity. In addition to the core growth there is also the growth in the number of threads a processor can run simultaneously to pump more computing through the cores.


Windows 7 supports up to 256 processors, and Linux supports a large number as well, but it is the applications that actually add value, not the operating system. What can be written that takes advantage of these advances may start to shift what is actually written.

One word as a focal point for change for 2008 - Collaboration

In an interview by InformationWeek in Mexico this week, they asked me a question that caught me a bit off-guard, at the time. It was something along the lines of: “If there is one word that you believe will be the focus of change for 2008, what would it be?” After I got past the deer in the headlights initial response, I came up with “Collaboration”. It can be applied at many levels to the changes that are underway.

 

At the cultural level, we’re all familiar with web 2.0 and the collaboration across organizations it supports. Wikinomics states the view of collaboration between organization, increases diversity of perspective enabling innovation and reaching objectives more quickly.

 

At the software level, the concept of SOA is based upon the collaboration between services, enabling clear separation between the interface and the underlying data, freeing up organizations to focus at a higher (more business oriented) level.

 

In the hardware space, we have multi-core. It enables multiple processing units to collaborate together on delivering value to the enterprise. To maximize the value delivered by working together, it will require new software and new approaches.


Companies need to be more agile, moving from viewing change as a periodic disruption of the status quo to accepting continuous change as the norm. Information technology (IT) has an important role to play, since it enables agility through collaboration. IT needs to collaborate with the rest of the enterprise in meeting the business objectives, until it fades into the business itself.


We’re in a race, and it is not a short sprint, nor is it a long distance endurance contest where the runner is able to alternate between coasting and sprinting to break the back of other runners. The high-stakes race to become the next-generation enterprise is a different kind of contest where the enterprise must consistently run faster and faster to keep pace with the 21st century’s accelerating rate of change. Collaboration at all these levels will be key.

Multi-core marches on

Last year I touched on the implications of Azul and Sun making strides around multi-core.

I saw this blog entry about Azul coming out with a 48 core chip next year. With all the gnashing of teeth going on about heat dissipation and limits of Moore's law, there are clearly other ways to get throughput.

This shift from sequential to parallel programming techniques will be disruptive to all the folks trained in sequential programming. We'll see new languages ... but with Model Driven and Service Oriented Architecture tools coming into vogue, the shift to parallel may be hidden to a large degree though. We can focus on delivering business value.

Labels: Multi-core| SOA
Search
Follow Us
About the Author(s)
  • Steve Simske is an HP Fellow and Director in the Printing and Content Delivery Lab in Hewlett-Packard Labs, and is the Director and Chief Technologist for the HP Labs Security Printing and Imaging program.
Labels
The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation