Lasse, On the rc_dec_speed PR, it appears to me the main change that might speed things up is this unsigned comparison? https://github.com/tukaani-project/xz-java/compare/master...rc_dec_speed#diff-8a1afbf1609c4b2d7813b299fce056f7ddd58a4a24ff02b01c2fdba38ff7fd0dL24-R23
Did you try out the Integer.compareUnsigned method as an alternative? https://docs.oracle.com/javase/8/docs/api/java/lang/Integer.html#compareUnsigned-int-int- My observation historically has been that moving to 64bit operations is slower. I will try to find some time to take a closer look at the PR you referenced. Brett On Thu, Jul 18, 2024 at 9:13 AM Lasse Collin <lasse.col...@tukaani.org> wrote: > > On 2024-03-07 Lasse Collin wrote: > > On 2024-03-05 Dennis Ens wrote: > > > > I hope 1.10 could be done in a month or two but I don't want to > > > > make any promises or serious predictions. Historically those > > > > haven't been accurate at all. > > > > > > I'll hope it's on the sooner side then. > > 1.10 is getting closer to being ready. I'd like to fix this-escape > warnings still as those are a sign of design errors. More info here: > > https://github.com/tukaani-project/xz-java/pull/18 > > I also created a branch rc_dec_speed but didn't create a PR for it. It > speeds up decompression on x86-64 with OpenJDK 22 roughly 4 %. I'm not > sure if it should be included in 1.10 in some form. If there is > interest, let's create a separate thread for it, or I can create a PR if > that is preferred. > > So now you and other people with Java knowledge can easily help because > especially the PR #18 doesn't need much knowledge of the project > internals. Thanks! > > I currently don't know what I should post to xz-devel and what to > GitHub. Many public communication channels makes things hard to follow. > > -- > Lasse Collin >
