[all-commits] [llvm/llvm-project] cd7e65: [flang] optimize array function calls using hlfir....

[jeanPerier via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: cd7e65398fbbd9642573013800dc3ae1e7307f82
      https://github.com/llvm/llvm-project/commit/cd7e65398fbbd9642573013800dc3ae1e7307f82
  Author: jeanPerier <jperier at nvidia.com>
  Date:   2024-12-03 (Tue, 03 Dec 2024)

  Changed paths:
    M flang/include/flang/Lower/ConvertCall.h
    M flang/include/flang/Optimizer/HLFIR/HLFIRDialect.h
    M flang/lib/Lower/ConvertCall.cpp
    M flang/lib/Lower/ConvertExpr.cpp
    M flang/lib/Optimizer/HLFIR/IR/HLFIRDialect.cpp
    M flang/lib/Optimizer/HLFIR/IR/HLFIROps.cpp
    M flang/test/HLFIR/order_assignments/where-scheduling.f90
    A flang/test/Lower/HLFIR/calls-array-results.f90
    M flang/test/Lower/HLFIR/where-nonelemental.f90
    M flang/test/Lower/explicit-interface-results-2.f90
    M flang/test/Lower/explicit-interface-results.f90
    M flang/test/Lower/forall/array-constructor.f90

  Log Message:
  -----------
  [flang] optimize array function calls using hlfir.eval_in_mem (#118070)

This patch encapsulate array function call lowering into
hlfir.eval_in_mem and allows directly evaluating the call into the LHS
when possible.

The conditions are: LHS is contiguous, not accessed inside the function,
it is not a whole allocatable, and the function results needs not to be
finalized. All these conditions are tested in the previous hlfir.eval_in_mem
optimization (#118069) that is leveraging the extension of getModRef to
handle function calls(#117164).

This yields a 25% speed-up on polyhedron channel2 benchmark (from 1min
to 45s measured on an X86-64 Zen 2).



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