[LLVMdev] Bringing back Andersen-like alias analysis

Duncan Sands baldrick at free.fr
Mon Feb 20 01:41:51 PST 2012


Hi Tristan,

> I am researching a more efficient approach to subset-based alias analysis in the
> style of Andersen's algorithm, and I noticed that there used to be an
> implementation of that for LLVM back in 2.6 (-anders-aa) which was removed
> because it was "not being actively maintained and had substantial problems". I'd
> be interested in knowing what was wrong with it (other than Andersen's algorithm
> being O(n^3)) so that I can avoid the same mistakes. Does anyone recall what the
> issues were?

it looks like the LLVMSlicer project may have resurrected Andersen's, or written
their own version: https://github.com/jirislaby/LLVMSlicer

It's a bit hard to tell because there doesn't seem to be any documentation
(hopefully good documentation will be added soon, since it won't be accepted
into the main LLVM project without it).

Ciao, Duncan.



More information about the llvm-dev mailing list