[all-commits] [llvm/llvm-project] 619720: [CodeGen] Limit number of analyzed predecessors

[Alexis Engelke via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 61972054f3fcaf59096799342bac9c93dd9aa432
      https://github.com/llvm/llvm-project/commit/61972054f3fcaf59096799342bac9c93dd9aa432
  Author: Alexis Engelke <engelke at in.tum.de>
  Date:   2025-06-20 (Fri, 20 Jun 2025)

  Changed paths:
    M llvm/lib/CodeGen/MachineBlockPlacement.cpp
    M llvm/test/CodeGen/RISCV/branch.ll

  Log Message:
  -----------
  [CodeGen] Limit number of analyzed predecessors

MachineBlockPlacement has quadratic runtime in the number of
predecessors: in some situation, for an edge, all predecessors of the
successor are considered.

Limit the number of considered predecessors to bound compile time for
large functions.

Pull Request: https://github.com/llvm/llvm-project/pull/142584



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