[all-commits] [llvm/llvm-project] af9b25: [RISCV] Optimize floating point scalar move and splat

[Kito Cheng via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: af9b25f9dbc261c3da42660181da3975e3a85d60
      https://github.com/llvm/llvm-project/commit/af9b25f9dbc261c3da42660181da3975e3a85d60
  Author: Kito Cheng <kito.cheng at sifive.com>
  Date:   2023-09-06 (Wed, 06 Sep 2023)

  Changed paths:
    M llvm/lib/Target/RISCV/RISCVInsertVSETVLI.cpp
    M llvm/test/CodeGen/RISCV/rvv/vsetvli-valid-elen-fp.ll

  Log Message:
  -----------
  [RISCV] Optimize floating point scalar move and splat

In D158086, we limit all floating point scalar move and splat can't fuse
vsetvli with different SEW, and this patch try to relax the constraint
as possible by introducing new SEW demand type:
SEWGreaterThanOrEqualAndLessThan64, that allow SEW fused with larger
SEW, but constraint it can't fused with SEW=64.

Reviewed By: rogfer01

Differential Revision: https://reviews.llvm.org/D158177