https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110945
--- Comment #8 from Jan Schultke <janschultke at googlemail dot com> --- (In reply to Jonathan Wakely from comment #4) > Please provide the testcase in a usable form, not just a link to an external > site (which uses its own custom benchmark macros). This is requested at > https://gcc.gnu.org/ Thanks, I will remember to do that in the future. > This is not equivalent to the other forms of copying in the benchmark, > because string::assign has to handle possible aliasing. It's valid to do > things like str.assign(str.data()+1, str.data()+2). >From what I could read in the `char_traits::move` code that presumably gets called, this function explicitly tests for overlap between the memory regions, and dispatches to cheap functions if possible. The input size was 8 MiB, so it is unlikely that the overhead from this overlap detection is contributing in any relevant capacity. Basically, due to this overlap testing, `assign` SHOULD be just as fast as other methods if there is no overlap (and in this case, there clearly is none). However, it is 14x slower, so something is off. Either I haven't followed the logic correctly, or there is a mistake in this dispatching logic which leads to much worse performance for .assign().
