[PATCH] D20789: Consecutive memory access in Loop Vectorizer - fixed and simplified

[Adam Nemet via llvm-commits]

anemet added inline comments.

================
Comment at: ../test/Transforms/LoopVectorize/consec_no_gep.ll:28-36
@@ +27,11 @@
+  %i.05 = phi i32 [ %inc, %for.body ], [ 0, %for.body.preheader ]
+  %from.addr.04 = phi float* [ %incdec.ptr, %for.body ], [ %from, %for.body.preheader ]
+  %to.addr.03 = phi float* [ %incdec.ptr1, %for.body ], [ %to, %for.body.preheader ]
+  %incdec.ptr = getelementptr inbounds float, float* %from.addr.04, i64 1
+  %0 = bitcast float* %from.addr.04 to i32*
+  %1 = load i32, i32* %0, align 4
+  %incdec.ptr1 = getelementptr inbounds float, float* %to.addr.03, i64 1
+  %2 = bitcast float* %to.addr.03 to i32*
+  store i32 %1, i32* %2, align 4
+  %inc = add nsw i32 %i.05, 1
+  %cmp = icmp slt i32 %inc, %len
----------------
Elena, are you planning to address this comment as well?


Repository:
  rL LLVM

http://reviews.llvm.org/D20789