[Bug tree-optimization/14741] graphite with loop blocking and interchanging doesn't optimize a matrix multiplication loop

[spop at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14741

--- Comment #29 from Sebastian Pop <spop at gcc dot gnu.org> ---
(In reply to Sebastian Pop from comment #28)
> delinearize the array accesses, and we don't have code to do that yet.

There is code to delinearize array accesses in LLVM now: it works on top of
SCEV, so it should be portable to GCC.  If somebody is interested to contribute
the delinearization to GCC, I can help. The LLVM code is in
llvm/lib/Analysis/ScalarEvolution.cpp SCEVAddRecExpr::delinearize()