# Friday, July 27, 2007

You know this is one of my hobbyhorses. But I didn't write this quote, and neither did Herb Sutter or Larry O'Brien:

And so the world is going to move more and more away from one CPU that is multiplexed to do everything to many CPUs, and perhaps specialty CPUs. This is not the world that the programmers target today. This kind of complexity was historically reserved only for the wizards who wrote the core operating system; or, in the world of supercomputing in science and engineering, people who had the ultimate requirement for computational performance built big machines like this and have used them to solve some of the world's tough computational problems. That was always a niche part of the industry.

This presages a change where the industry at large, the whole concept of applications, will ultimately have to be restructured in order to think about how to take advantage of these machines, because they won't just get faster every year. They'll get more powerful, but in fact only if you're able to master these problems of concurrency and complexity.

The concurrency is a real challenge, because the way the industry has grown up writing software - the languages that we chose, the model of synchronization and orchestration, are actually not things that lend themselves toward either exposing parallelism or allowing large-scale composition of big systems and it's in part why we and everybody else, as the software grows in scale, you know, deal to a greater and greater degree with the difficulty of perfecting the software, making it absolutely secure, being able to predict every aspect of its operation. And so today we face the dual challenge of having the prospect of meeting even bigger, more sophisticated pieces of software to do the powerful things that we want, and to do it in an environment where to get that performance at the client on an individual application will require the mastery of parallelism.

This is Microsoft's Chief Research & Strategy Officer, folks. And he says what I say: concurrency is hard, and the future is concurrent. I know we all get by in this crazy churning world of constant new releases by ignoring stuff, but you can't ignore this.


Friday, July 27, 2007 10:40:15 PM (Eastern Daylight Time, UTC-04:00)  #