How big is your monitor?  And how many do you have?

Posted by Webmaster on May 11th 2012 | Discuss
Tags:

There’s a lot of assumptions running around in this PC industry, based on hearsay and tidbits of data. We just finished running some tests on Nvidia’s new dual-GPU AIB, the GTX 690 (see p. 11). Nvidia wasn’t very happy with our first results, and told us we were being CPU-bound by doing the tests at HD. Aside from the counter-intuitive comments (after all, isn’t that exactly what a GPU is supposed to do—eliminate CPU limitations?), it also raises the question about what the potential buyers of this $999 AIB have on their desks. I suppose it is reasonable to expect that a limited production super-high-end AIB (e.g., Asus only made 1,000 of their killer Mars II dual GPU that sold for $1,500) would only be bought by users with big bucks and big screens.

We looked at three sources to see if we could get a comfortable feeling about monitor size in use.

We looked at Steam’s data. We contacted StatCounter Global Stats, who monitor browsers (for web developers) and sniff the systems behind them. And we checked with Futuremark.

The results are somewhere between interesting and amazing, and prove once again how fragile conclusions about the user base can be depending on your sample size and demographics.

Futuremark’s data, which is reported by users of their benchmarks, shows a steady increase in larger screen resolutions. Steam also shows a similar trend, as does StatCounter—but the resolutions differ widely.

Futuremark has a report from over 176,000 users (every month).

StatCounter uses javascript to obtain the screen resolution of the devices they track. Their screen resolution statistics specifically do not include mobile devices due to the vast array of different mobile screen resolutions. StatCounter reports that 1366 x 768 has become the most popular screen resolution worldwide, having overtaken 1024 x 768. We were surprised that 2560 x 1600 doesn’t even show up in the survey, which is probably due to the large sample base and the nature of the users.

Since the company began its tracking of screen resolutions in March 2009 as a free service to developers and other users, StatCounter reports that 1024 x 768 has been the dominant screen size globally on the web.

“The data reflects a continuing trend of users moving to larger screen resolution sizes,” commented Aodhan Cullen, CEO, StatCounter. “The screen resolution size people are using is a critical factor for developers when it comes to web design, particularly in the case of fixed-width web pages.”

In January 2012, StatCounter’s ­global sample of screens used for popular browsers consisted of more than 18.1 billion hits. The ten countries with the largest individual sample sizes are listed below:

  • 4.4 billion – United States
  • 1.1 billion – Turkey
  • 925 million – India
  • 893 million – Brazil
  • 872 million – United Kingdom
  • 657 million – Canada
  • 568 million – Germany
  • 516 million – Thailand
  • 450 million – Spain
  • 408 million – Italy
SCREEN RESOLUTION: Comparison of StatCounter’s population and Steam and Futuremark (Source: Jon Peddie Research)

All of this leads us to the conclusion that StatCounter represents the every­day user. Steam is used by a broad range of people, but they are all gamers. Future­mark, on the other hand, represents performance/enthusiast users. These are people who are comparing the performance of their machines against others to see if they can improve performance.

What about the kind of people who actually do go out and buy a dual-GPU monster AIB? What resolutions do they run anyway? Checking Futuremark’s database for users who had AMD Radeon HD 6990, the percentage of users running 2560 is much higher, about 7%, but 1080p is still the most common reso­lution with over half of the users running that. With GTX 590 added. Numbers are similar to Radeon HD 6990 users, but there are about a percentage point more users who run 2560-wide resolutions.

The bottom line is you can’t make conclusions about screen resolution without a lot of qualifications about the user.

Discuss this editorial