Hi, This patch series improves the costing model in the AArch64 backend to match a number of new idioms.
This patch is a combination of a series I had been working on, with the cost-model for XGene-1 proposed by Philipp Tomsich. ( http://gcc.gnu.org/ml/gcc-patches/2014-02/msg01084.html ) Where sensible I have integrated the idiom matching in Philipp's with my own work, though there were cases which were redundant, or could be folded to reduce code duplication. There were other cases where the code suggested XGene-1 would benefit from special-case handling. Without documentation for the XGene-1 I can't cater for these special cases, and I have not tried to do so here. The patch series has been bootstrapped natively on aarch64-none-elf, and has been through aarch64-none-elf runs with no issue. Is this OK for stage-1? Thanks, James
