I read your pass “Pre-increment preparation pass” on LLVM-commits mail-list.
I appreciate your idea in the talking with Andrew.
Recently, I plan to implement a relative pre-increment of memory ops of my back-end in the
stage of LSR(LoopStrengthReduce).
And now, do you have update your thought and code of this pass?
Any suggestion will be great.
Thank you very much in advance.