[llvm] [DA] use NSW arithmetic (PR #116632)

[Michael Kruse via llvm-commits]

================
@@ -3116,11 +3116,9 @@ const SCEV *DependenceInfo::addToCoefficient(const SCEV *Expr,
                                              const Loop *TargetLoop,
                                              const SCEV *Value) const {
   const SCEVAddRecExpr *AddRec = dyn_cast<SCEVAddRecExpr>(Expr);
-  if (!AddRec) // create a new addRec
-    return SE->getAddRecExpr(Expr,
-                             Value,
-                             TargetLoop,
-                             SCEV::FlagAnyWrap); // Worst case, with no info.
+  if (!AddRec)
+    return SE->getAddRecExpr(Expr, Value, TargetLoop, SCEV::FlagNSW);
----------------
Meinersbur wrote:

> NoWrapFlags is set to FlagNSW for the evolution function to be linear.

The NSW flag doesn't make it linear in the mathematical sense, which is impossible on a finite domain. The SCEV still has a type limited to some integer bitwidth, and ScalarEvolution makes use of the fact, e.g. when simplifying expressions.

But if the math behind MIV subscripts are based on linear(/affine) functions on $\mathbb{R}^n \to \mathbb{R}$ with unbounded range (red race car book, p129) means that any wrapping behaviour of $a^+l$ and $a^-u$ would also not be correct indicated by the "bad subscript classification", maybe even less so.

https://github.com/llvm/llvm-project/pull/116632