Commit fd57d66

committed

Vectorize: Support fminimumnum and fmaximumnum

Support auto-vectorize for fminimum_num and fmaximum_num. For ARM64 with SVE, scalable vector cannot support yet, and For RISCV Vector, scalable vector works well now.

1 parent fff8f03 commit fd57d66Copy full SHA for fd57d66

4 files changed

+426

-2

lines changed

clang/test/CodeGen
- fminimum-num-autovec.c
llvm
- include/llvm/CodeGen
  - BasicTTIImpl.h
- lib
  - Analysis
    - VectorUtils.cpp
  - Target/RISCV
    - RISCVISelLowering.cpp

4 files changed

+426

-2

lines changed

Comments

(0)

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Commit fd57d66

4 files changed

4 files changed

File tree

4 files changed

4 files changed

0 commit comments