Date: Thu, 23 Jul 2020 18:00:01 +0000 From: Colin Percival <cperciva@tarsnap.com> To: Doug Moore <dougm@FreeBSD.org>, src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: Please include benchmarks in commit messages (was Re: svn commit: r363450 - in head: share/man/man3 sys/sys) Message-ID: <010001737cd501a6-34cc6066-95c3-4724-a5f7-7cb00193e223-000000@email.amazonses.com> In-Reply-To: <202007231716.06NHGLNF020979@repo.freebsd.org> References: <202007231716.06NHGLNF020979@repo.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
[Picking this as an example of a general trend...] On 2020-07-23 10:16, Doug Moore wrote: > [...] > Testing has shown that for the cases where red-black trees do worst, > wavl trees better balance leads to faster lookups, so that if lookups > outnumber insertions by a nontrivial amount, lookup time saved exceeds > the extra cost of balancing. I'd like to request that commits like this include some benchmarks. In the review of these particular changes I see several microbenchmarks for a variety of use cases; it could be helpful to future code spelunkers if these were summarized in the commit message (e.g. "lookup-heavy microbenchmarks saw a X% speedup while write-heavy microbenchmarks saw a Y% slowdown"). It would also be nice to see a macrobenchmark (e.g. buildworld), even (and perhaps *especially*) if the result is "no difference was observed". Again, this is a general request -- this is just the latest in a long series of commits I've seen by many people which have sounded like they could have performance impacts but have not included any quantitative results. -- Colin Percival Security Officer Emeritus, FreeBSD | The power to serve Founder, Tarsnap | www.tarsnap.com | Online backups for the truly paranoid
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?010001737cd501a6-34cc6066-95c3-4724-a5f7-7cb00193e223-000000>