[all-commits] [llvm/llvm-project] e2202b: [AArch64] Update costs for scalarizing i64->f32 in...

[David Green via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: e2202b944b56039498f2445a1c4ff984a6a303ee
      https://github.com/llvm/llvm-project/commit/e2202b944b56039498f2445a1c4ff984a6a303ee
  Author: David Green <david.green at arm.com>
  Date:   2025-03-26 (Wed, 26 Mar 2025)

  Changed paths:
    M llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
    M llvm/test/Analysis/CostModel/AArch64/cast.ll
    M llvm/test/Analysis/CostModel/AArch64/no-sve-no-neon.ll
    M llvm/test/Analysis/CostModel/AArch64/sve-cast.ll

  Log Message:
  -----------
  [AArch64] Update costs for scalarizing i64->f32 int_to_fp. (#132366)

After #130665 these operations are scalarized to avoid
double-rounding. This updates the cost model to match.

In the future we might be able to use SVE instructions to help, but for
the moment the costs should be higher. Costsize and Latency costs are
not yet expected to be accurate. The vector insert/extract will use the
cost of VectorInsertExtractBaseCost (2 by default).



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications