[PATCH] D74691: [Attributor] Detect possibly unbounded cycles in functions

[Stefanos Baziotis via Phabricator via llvm-commits]

baziotis added a comment.

In D74691#1892538 <https://reviews.llvm.org/D74691#1892538>, @omarahmed wrote:

> In D74691#1892476 <https://reviews.llvm.org/D74691#1892476>, @baziotis wrote:
>
> > In D74691#1892447 <https://reviews.llvm.org/D74691#1892447>, @omarahmed wrote:
> >
> > > I am sorry but I can't get what the test that the approach of moving on the SCCs fails in as I have tried to test it a bit with tests like that and printed the SCCs that it gets
> > >
> > > https://i.imgur.com/fpZhpST.png
> > >  1 - entry
> > >  2- condition - body
> > >  3- return
> > >  4 - l4
> > >
> > > https://i.imgur.com/UJJ3aH0.png
> > >  1- entry
> > >  2- l1 - l2 - l3 - l4 - l5
> > >  3- return
> > >  4- l6 - l7 - l8
> > >
> > > https://i.imgur.com/92NS6rp.png
> > >  1- entry
> > >  2- l1 - l2 - l3 - l4 - l5 - l6 - l7
> > >  3- return
> > >
> > > and in all of them it reached the cycle and outputs that the loop does not have a max trip count
> >
> >
> > Sorry, but I didn't understand your comment very much. The code //I// wrote certainly can't detect max trip count.
> >  It only correctly (hopefully) finds if there's a cycle or not. The second modification also found the size of the cycle.
> >  Did //you// write any code that used `LI` ? In that case, as I said earlier, I have some algorithm in mind that uses `LI`
> >  but I can't seem to be able to get `LI` (i.e. it's always `nullptr`). So, probably the same will be true for you
> >  even if your code is correct. Hopefully we can get an answer about why that happens.
>
>
> I mean the code you suggested that loops on SCCs using SCCiter then it appeared that tarjan only detects maximal SCCs, the one that uses tarjan algorithm :)


With that code the reason you don't find any max trip count as I said is that you probably don't have `LoopInfo`. Running the examples,
I could not get `LoopInfo` in any way, even with `-loops -analyze`. That's a problem that we have to fix so that the new algorithm can be improved
with `LoopInfo` as well.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D74691/new/

https://reviews.llvm.org/D74691