[PATCH] D21909: [DSE]Enhance shorthening MemIntrinsic based on OverlapIntervals

[Jun Bum Lim via llvm-commits]

junbuml added inline comments.

================
Comment at: lib/Transforms/Scalar/DeadStoreElimination.cpp:834
@@ +833,3 @@
+  int64_t NewLength = IsOverwriteEnd
+                          ? LaterOffset - EarlierOffset
+                          : EarlierSize - (LaterOffset - EarlierOffset);
----------------
mcrosier wrote:
> Is this clang-formatted?  Spacing looks odd.
Yes, I got this from clang-format.

================
Comment at: lib/Transforms/Scalar/DeadStoreElimination.cpp:1052
@@ -937,9 +1051,3 @@
                      isShortenableAtTheBeginning(DepWrite)))) {
-          // TODO: base this on the target vector size so that if the earlier
-          // store was too small to get vector writes anyway then its likely
-          // a good idea to shorten it
-          // Power of 2 vector writes are probably always a bad idea to optimize
-          // as any store/memset/memcpy is likely using vector instructions so
-          // shortening it to not vector size is likely to be slower
-          MemIntrinsic *DepIntrinsic = cast<MemIntrinsic>(DepWrite);
-          unsigned DepWriteAlign = DepIntrinsic->getAlignment();
+          assert(!EnablePartialOverwriteTracking && "Do not expect to perform "
+                                                    "when partial-overwrite "
----------------
mcrosier wrote:
> Is this how clang-format formatted the string?
Yes, this is what clang-format suggested.


https://reviews.llvm.org/D21909