Sebastian Pop wrote: > Diego, can you be more specific about what parts you think have to be > lowered more from GIMPLE?
It's something we've discussed on and off for a couple of years. One idea is to expose in GIMPLE target features like word size, pointer arithmetic, etc. It's not something that would require changes to GIMPLE (in principle), it would just expose these features and make GIMPLE a sort of high-level RTL. I have not really explored the idea yet, so I can't say if it's a good direction. The intuition is that we could reuse existing SSA optimizers for the initial set of RTL cleanups. > Together with Richard Guenther, we're planing to send a proposal for > having higher level vector constructs in GIMPLE. Looking forward to that. I can't tell you whether I like the idea or not, but I'm ready to be convinced.