Utility of Multicore Chips?

I have to admit that I'm a skeptic of multicore chips. Even though they're part of many CPUs shipping today, it's just not clear to me what problem they solve. We've already had marginal returns on CPU performance, in terms of human productivity. The bottleneck just isn't the microprocessor anymore.

Don Knuth also sees some challenges for multicore:

Let me put it this way: During the past 50 years, I’ve written well over a thousand programs, many of which have substantial size. I can’t think of even five of those programs that would have been enhanced noticeably by parallelism or multithreading. Surely, for example, multiple processors are no help to TeX.


On the other hand, multicore might represent an opportunity. One of the trends in research this past decade has been novel ways of "wasting" CPU to enable other desirable properties, such as security and usability. All we need now is a clearer path for making this happen.

Comments

Post a Comment

Popular posts from this blog

How to Fix a Jammed Toyota Camry Trunk

This problem needs a higher pagerank, so I figured I would post the solution here. If your Toyota Camry trunk won't open, one possible reason is that it is set to valet mode. Valet mode means that you cannot open the trunk using the release lever inside the car. To set valet mode, you put the key into the trunk lock and turn it counterclockwise. You will know that your trunk is in valet mode if the lock is horizontal rather than vertical, and if you cannot open the trunk using the lever near the driver's seat. Of course, a problem is that sometimes the Camry can get stuck in valet mode, such that you can't use your key to get out of it. (You can see how I spent part of my Sunday morning ...) The solution turns out to be WD-40 . Spray some WD-40 on your key and on the lock. Put the key in, and jiggle it around, and happiness ensues. From an interaction design perspective, it sort of makes sense to have a valet mode. After all, the point of having a valet key is to limit the
Read more

Web 2.0 and Research

I've been chatting with many of my friends and colleagues about an issue that's been bugging me for a while, namely whether academic research has any role to play in the emerging Web 2.0 . I've been slowly coming to the conclusion that the answer is not much. I had a similar discussion with other researchers at HotMobile a few weeks ago. When the web first came out, pretty much every systems researcher ignored it because it was so ugly. The web was not very sophisticated in terms of distributed systems, HTTP lacked elegance, HTML conflated many different ideas, and so on. There were also not any really new ideas with the web, as evidenced by the fact that Tim Berners-Lee 's first paper on the Web was (probably rightfully) rejected from an ACM conference on hypertext. I'm sure one thing that really irked researchers about the nascent web was that it completely ignored the large body of work in hypertext and distributed systems that had preceded it. Even in 1997, as
Read more

NYTimes: It Takes a Cyber Village to Catch an Auto Thief

I like this idea of "open-source crime solving", as it reminds me a lot of PhishTank and CastleCops . It is, however, an idea that is fraught with issues of trust, reliability, and vigilantism. Online auto forums have helped unravel crimes before. Two years ago, a detective in Los Angeles used the forum on FreshAlloy.com, a Nissan enthusiast site, to track down victims of an elaborate fraud scheme. (That case, too, involved Nissan Skylines.) The Beyond.ca site had also played a role in earlier cases of what might be called open-source crime solving. A year ago one of its members saw a hit-and-run accident a block in front of him, said Shelton Kwan, who co-founded the site with his cousin Ken Chan in 2002. “He took pictures. And the guy who got hit was another member of ours.” NYTimes Link
Read more