Architectures like Mips are very limited when it comes to addressing modes. Therefore, the expected behavior would be that, for the BASE + OFFSET addressing mode, complexity is lower, while, for more complex addressing modes (e.g. BASE + INDEX << SCALE), which are not supported, complexity is higher. Currently, the complexity calculation algorithm bails out if BASE + INDEX addressing mode is not supported by the target architecture, resuling in 0-complexities for all candidates, which leads to non-optimal candidate selection, especially in scenarios where there are multiple nested loops.
Additionally, when bumping up the register pressure cost, the number of invariants should also be considered, in addition to the number of candidates. Dimitrije Milosevic (2): ivopts: Revert computation of address cost complexity. ivopts: Consider number of invariants when calculating register pressure. gcc/tree-ssa-address.cc | 2 +- gcc/tree-ssa-address.h | 2 + gcc/tree-ssa-loop-ivopts.cc | 220 +++++++++++++++++++++++++++++++++++++++++--- 3 files changed, 210 insertions(+), 14 deletions(-) --- 2.25.1
