The first thing that I found interesting was that, asked why they wouldn't use customizable chips on in their next project, respondents overwhelmingly (68 percent) said they didn't need the programmable functionality. The number of people who gave this response dwarfed the number who said PLDs were too expensive, power hungry or too difficult to use.
There was some debate about these numbers. It was pointed out during the presentation that this year was the first time respondents were offered the "don't need this functionality" choice. One of the presenters, Michael Barr, president and CTO of Netrino pointed out that this likely skewed the numbers; in past years people were forced to give another answer, so they likely blamed it on one or more of the old standbys, cost, power and difficulty.
In response to another question, 55 percent said there weren't using FPGAs in their current designs, while 45 percent said they were.
And finally, take a look at this chart below (click to enlarge). It shows that a lot more people said they would consider FPGAs from the various vendors for their next project than our actually using them in their current designs. At first glance, this would seem to be great news for FPGAs (more users). But as Rich Nass pointed out during the presentation, you have to remember that people might be willing to consider more options, but they are still going to choose only one. So they longer bars on the bottom might just be attributed to more test drives.
Click on image to enlarge.
I would encourage anyone who has not done so to check out this presentation on demand on TechOnline. The FPGA stuff is towards the end, but there is plenty of other interesting (and sometimes very surprising) information.