On Fri, 17 Feb 2023 15:24:26 GMT, Raffaello Giulietti <rgiulie...@openjdk.org> wrote:
>>> After making sure that `intVal` is even, and before attempting a division >>> by a power of 10, it might help to check if 5 divides `intVal` in the first >>> place. If it doesn't, there no point in performing the division. >>> >>> It can be shown that 5 divides `intVal` if and only if it divides the >>> `long` sum of all `int`s in the `mag` array of `intVal`. >>> >>> I didn't try out if this contributes to improve the overall performance, >>> but it might be worth giving a try. >> >> Unfortunately, it seems this division by 5 doesn't give an obvious >> performance improvement. Maybe previous lowest bit check has filtered some >> cases. > >> > After making sure that `intVal` is even, and before attempting a division >> > by a power of 10, it might help to check if 5 divides `intVal` in the >> > first place. If it doesn't, there no point in performing the division. >> > It can be shown that 5 divides `intVal` if and only if it divides the >> > `long` sum of all `int`s in the `mag` array of `intVal`. >> > I didn't try out if this contributes to improve the overall performance, >> > but it might be worth giving a try. >> >> Unfortunately, it seems this division by 5 doesn't give an obvious >> performance improvement. Maybe previous lowest bit check has filtered some >> cases. > > Thanks for trying out. @rgiulietti Hi, could you please review this pr ------------- PR: https://git.openjdk.org/jdk/pull/12509
