[PATCH] D93855: [RISCV] Define vector widening type-convert intrinsic.

[Craig Topper via Phabricator via llvm-commits]

craig.topper added a comment.

Do these instructions need earlyclobber?



================
Comment at: llvm/include/llvm/IR/IntrinsicsRISCV.td:393
+  // For Conversion unary operations.
+  // For destination vector type is the same as first source vector.
+  // Input: (vector_in, vl)
----------------
The destination vector type isn't the same as the first source vector so this comment seems wrong


================
Comment at: llvm/include/llvm/IR/IntrinsicsRISCV.td:400
+  // For Conversion unary operations with mask.
+  // For destination vector type is the same as first source vector.
+  // Input: (maskedoff, vector_in, mask, vl)
----------------
Same here


================
Comment at: llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td:261
+defset list<VTypeInfoToWide> AllWidenableIntToFloatVectors = {
+  def : VTypeInfoToWide<VI16MF4,  VI32MF2>;
+  def : VTypeInfoToWide<VI16MF2,  VI32M1>;
----------------
Why can't we just list the float types as the second argument here instead of needing use GetFloatVTypeInfo?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D93855/new/

https://reviews.llvm.org/D93855