As a technologist
and technology journalist
for over twenty years, I've been a witness or participant to many different game-changing technologies. Some end up changing the way we look at computing, while others just seem to fizzle out and disappear after a while.
I have little doubt that GPU computing will fall into the former category. That assessment is based on the fact that the performance of GPUs on many important scientific and engineering problems overshadows just about any other method of computation, save a supercomputer. But GPUs are typically less than the cost of CPUs, and have a place in any engineer's desktop workstation.
There are huge hurdles to overcome. The largest are probably the availability of good software development tools, and the motivation of engineering software vendors to build applications that can take advantage of the floating point computation abilities of the GPU.
This year, I'll be presenting at the GPU Technology Conference
, discussing why GPU computing is changing the face of engineering computing, and how software vendors are getting innovative about enabling engineering code to run on those processors. I'll provide data concerning the adoption rate of GPUs among engineering users and discuss how both engineers and software developers can make greater use of GPUs in solving computationally-intensive problems.
My session, titled GPU Computing and a Revolution in Design Engineering
,takes place Tuesday, September, 21st, 11:00 - 11:50am. Hope to see you there.