This is a community focused on architecture and/or microarchitecture research with the aim of impacting future generation GPU computing hardware.  An important part of this research is simulation tools that can model future generation hardware.  These simulation tools may also give application developers insight into the performance of their application.

Group members are invited to post links to publications, calls for papers, and other research resources on any of the following topics:

• GPU computing related microarchitecture, architecture, or programming model proposals
• GPU computing benchmarks (or important applications) for evaluating future generation architectures
• Simulation tools and methodologies

In my own research group we have developed a simulator called GPGPU-Sim that runs CUDA and OpenCL applications (more information here).  We have used this simulator to explore microarchitecture optimizations for control flow (branch divergence) and memory system hardware optimizations (more information here).

 - Tor M. Aamodt, University of British Columbia