[PATCH] D158716: [RFC][LV] VPlan-based cost model

[ShihPo Hung via Phabricator via llvm-commits]

arcbbb added inline comments.


================
Comment at: llvm/lib/Transforms/Vectorize/LoopVectorize.cpp:8700
+        C.first = VPCM.expectedCost(CostVF);
+        C.second = true;
+      } else
----------------
wangpc wrote:
> Is it always true even for scalar VF? Why can't we have the same signature as `CM.expectedCost`?
VectorizationCostTy is not exported in the header, so I change the interface trying to decouple the dependency.


================
Comment at: llvm/lib/Transforms/Vectorize/VPlanCostModel.h:30
+
+  /// Return cost of the VPlan for a given \p RVL
+  InstructionCost expectedCost(ElementCount VF);
----------------
wangpc wrote:
> `RVL`->`VF`?
Fixed. Thanks!


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D158716/new/

https://reviews.llvm.org/D158716