I have Google Alerts send me news on three topics:
- "Eugene Mirman" - my brother is a comedian, and I like to keep tabs on his touring schedule and TV appearances;
- "Guns N' Roses" - still keeping hope alive that Slash and Axl will reconcile, and the band will reunite;
- And of course, "MULTICORE" - what's new in the multicore world, advances on the silicon side of things, as well as the programming of multicore chips.
Jonathan Ericson's great column, "Parallelism: It's Everywhere, It's Everywhere" in the May 2nd Dr. Dobb's Report got me thinking. In it, he reports on Stanford University's formation of a Pervasive Parallelism Lab, and reflects on the multicore buzz:
So is all the buzz about parallelism the real deal or is it just more PR? Well, Bill Dally, chair of the computer science department at Stanford, thinks it's the real deal: "Parallel programming is perhaps the largest problem in computer science today and is the major obstacle to the continued scaling of computing performance that has fueled the computing industry, and several related industries, for the last 40 years."
Ten minutes later, I received the day's Google Alerts regarding scores of stories dealing with multicore chips, and the various efforts dealing with the programming of these chips. And an hour later, Yahoo! News told me about a whole bunch more. The drumbeat is unmistakable: for everyone involved in computing - vendors of chips, hardware, system and tools software, developers of application software, etc. - the issues around multicore are keeping more than a few of us up at night. The avalanche of news, though, feels like a relatively recent phenomenon, but I wasn't sure - so decided to look. The results surprised me.
Multicore in the News
As a proxy for news coverage, I checked out a few print and on-line sources:
- Dow Jones' Factiva service, great for looking at news coverage in printed publications
- Google News, to capture online news features
- Google Blogs, as a proxy for the blogosphere's attention to multicore
First, here's the Factiva data for the phrase "multicore". An order of magnitude change in coverage in just a couple years, with the "hockey stick" starting sometime around 2004 - my guess is that this related to the first mainstream multicore chips, dual-core x86, hitting the market.
To better focus on stories dealing with the multicore programming, I decided to play around with the search criteria a bit more:
- Volume of printed news stories that have both "multicore" and "programming" in the headline and/or 1st paragraph;
- Volume of online mentions of "multicore programming" in Google News
- And for fun, I searched Factiva for "global warming" mentions
I looked at a bunch of different search terms, and the results are fairly consistent. Here are the results (note that the vertical axis is a log scale).
Here's a Google News timeline for the multicore programming search term:
Couple observations:
- The growth of coverage for all the search terms appears exponential;
- Before 2005, there's no record in the Factiva database of a story where "multicore" and "programming" were in the headline and/or lead paragraph;
- The growth rate for the multicore-related searches is outpacing the growth rate for global warming coverage!!
Multicore in the Blogosphere
This is echoed in the blogosphere as well. Although Google Blog Search is still in beta (and the results change a bit from one run to the next), there's a couple eye-opening stats:
- In all of 2006, there were 329 blog posts dealing with multicore programming;
- Last month alone (April 2008), there 434 blog posts on the topic.
Wow! Exponential growth in coverage, perhaps correlated with an exponential growth in the concerns around the topic. The other thing I find curious is the timing: the news coverage "woke up" just as the first multicore processors were hitting the market. But that roadmap was clear years before! And yet, relatively little attention paid, and arguably modest progress in the way of tools to solve the multicore programming challenge.
Calling Al Gore...
Clearly, both global warming and multicore programming are getting a lot of attention these days. And hopefully, there will be progress on both, soon. Al Gore has done a great job alerting the world to global warming, and my hope is that he will now help spread the word about the multicore software challenge. Otherwise, without progress on these two key issues facing the globe, by the year 2022 there will be hundreds of millions of news stories on both topics, and multicore programming will overtake global warming in terms of coverage:
What do YOU Think?
- When did you start worrying about multicore programming?
- Is it time for Al Gore to help the multicore community?