General Cleanups
Website/Branding:
- [x] polygeist.mit.edu @wsmoses [almost done, pending https and some fastcgi fixes].
- [ ] Include relevant content [initially forked enzyme.mit.edu]
- [ ] Logo!
- [x] Rename repo @wsmoses
- [ ] Merge in Polymer repo @kumasento
Upstreaming:
- [x] Loop Naturalization [form scf.while from cfg]
- [x] Loop canonicalizaation [e.g. scf.while to scf.for]
- [x] Remaining SCF canonicalizations [like if]
- [x] Integer math canoncalization
- [x] AffineCFG
Features/Improvements:
- [x] Struct Types @wsmoses
- [ ] Improved parallelization [if you have time @ftynse ]
- [ ] Splitting Heuristic [@kumasento ]
- [ ] Vectorization / Unrolling at MLIR level [? @chelini ]
Do we also plan to rename the tool from mlir-clang to polygeist?
Maybe we could name it mlir-cfe based on its actual purpose IMHO
On Sat, 1 May 2021 at 18:30, lorenzo chelini @.***> wrote:
Do we also plan to rename the tool from mlir-clang to polygeist?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/wsmoses/Polygeist/issues/48#issuecomment-830666272, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACC42R33OV63YAFPJ5S6ELLTLQ3EXANCNFSM43VHPIIQ .
I'm not opposed to renaming it, though I do think that mlir-clang is more descriptive than mlir-cfe. Polygeist on own seems fine, though I wonder if something shorter could be done, maybe polyg / polyg++
cfe, that's a name I haven't heard in a long time.
I prefer mlir-clang. It makes clear what exactly the tool is, doesn't use an obscure outdated acronym (also note that cfe was the subproject name for clang, but the binary was named clang) and doesn't pretend it is the entire polygeist project (which it is not without polymer parts).
On Sat, May 1, 2021, 20:07 William Moses @.***> wrote:
I'm not opposed to renaming it, though I do think that mlir-clang is more descriptive than mlir-cfe. Polygeist on own seems fine, though I wonder if something shorter could be done, maybe polyg / polyg++
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/wsmoses/Polygeist/issues/48#issuecomment-830671516, or unsubscribe https://github.com/notifications/unsubscribe-auth/AALRG27IYN4XJGYRB5LSW73TLQ7OLANCNFSM43VHPIIQ .