On 08/04/2014 03:30 PM, Matt Arsenault wrote: > This needs a test. I’m guessing this is a branch on a constant? I’ve encountered problems like this from those in all of the GPU CFG passes Or a branch on an i1 function argument