gilbo

Simple plan: copy the CUDA source. That'll avoid adding weird build dependencies.

This requires having some kind of functional call abstraction that's consistent, along with function types of some sort in the typechecker. The function call abstraction is complicated because there are...

We would like to support some degree of profiling even just for our own use. Things like: - memory consumption - count of function launches - timing of how long...

Goal: We would like the Legion Spy graphs to have useful colorings. Underlying Issue: We re-use the same trampoline tasks for more or less all task launches right now, so...

We want to have a way to throw type-checking errors when certain accesses use non-constant values. We want this for Affine-Indexing, and also for indexing into vectors and matrices. In...

I need to go through the gpu_utils.t file and make sure that we're actually using the most specific atomic and reduction instructions available. This is definitely not the case currently.

Should we use tiny regions or futures? Problem with tiny regions: Does Legion support this with reasonable performance or will it overload the system? Problem with futures: Futures don't have...

We need to add back in Insert/Delete support. It's deprecated everywhere and not supported under Legion at all. See #37 for getting insert/delete working on GPUs

Right now, any error that happens while a kernel is running will cause a stack dump to report the error as originating within our compiler. That's bad for us. There...