[llvm] [SLPVectorizer] Widen strided loads. (PR #153074)

[Alexey Bataev via llvm-commits]

================
@@ -2214,6 +2228,35 @@ class BoUpSLP {
   ///       may not be necessary.
   bool isLoadCombineCandidate(ArrayRef<Value *> Stores) const;
 
+  /// Suppose we are given pointers of the form: %b + x * %s + y * %c
+  /// where %c is constant. Check if the pointers can be rearranged as follows:
+  ///  %b + 0 * %s + 0
+  ///  %b + 0 * %s + 1
+  ///  %b + 0 * %s + 2
+  ///  ...
+  ///  %b + 0 * %s + w
+  ///
+  ///  %b + 1 * %s + 0
+  ///  %b + 1 * %s + 1
+  ///  %b + 1 * %s + 2
+  ///  ...
+  ///  %b + 1 * %s + w
+  ///  ...
+  ///
+  ///  If the pointers can be rearanged in the above pattern, it means that the
+  ///  memory can be accessed with a strided loads of width `w` and stride `%s`.
+  bool analyzeRtStrideCandidate(ArrayRef<Value *> PointerOps, Type *ElemTy,
+                                Align CommonAlignment,
+                                SmallVectorImpl<unsigned> &SortedIndices,
+                                StridedPtrInfo *SPtrInfo) const;
----------------
alexey-bataev wrote:

Better to pass it always and avoid checking for nullptr.

https://github.com/llvm/llvm-project/pull/153074