[PATCH] D11977: Use modulo semantic to generate non-integer-overflow assumptions

[Johannes Doerfert via llvm-commits]

jdoerfert added a comment.

In http://reviews.llvm.org/D11977#222725, @grosser wrote:

> I did not had a detailed look yet, but wondered what the LNT results for
>  this patch looked like? Did this pass all tests and did we see a compile
>  time impact?


It should pass all lnt tests but I'll run lnt on this version again tonight.

The compile time impact is severe for <5 tests and "ok" for the rest. Most tests do not show worse compile time actually. However, I am still working on some simplifications that will reduce compile time as well as some exit conditions that we can place in order to bail out before we spend to much time on a benchmark.


http://reviews.llvm.org/D11977