Chris will report a few highlights fresh from the GPU Technology Conference in San Jose the week before. Jonathan will outline the fundamentals of GPU programming and accompany his presentation with some visually interesting examples that drive home a few of the issues that require attention.
Call[masked] if outside front door is locked ...
We discussed in the Jan 31 MeetUp that the meeting after next would be a review of the OpenCL and CUDA frameworks needed to get code going.
Meeting objective would be to have a decent understanding of the parts needed to set up a program in CUDA/OpenCL so that next meeting we could take on some actual parallel development/coding.
If you're bringing a CUDA-enabled laptop (recommended), download http://support.scinet.utoronto.... and make sure you can compile and run meetup-cuda/example1/example1.cu with the NVIDIA compilers.
If you don't have a CUDA-enabled laptop, we'll try something a little different. First, download http://support.scinet.utoronto.... . Then get a free account at http://www.nclab.com ; once you've logged in (doesn't work for me under safari, works fine under Chrome), go under Computing -> CUDA, delete all the stuff in the text box and replace it via copy-and-paste with the testpattern.py code; press "Play" and make sure that works. It's a different environment (python), but probably the best way to play with CUDA code if you don't have a CUDA laptop handy.