[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <fad03db9-342d-79b8-9b48-8865fb9f8d5f@ispras.ru>
Date: Sun, 7 Dec 2025 22:44:00 +0300 (MSK)
From: Alexander Monakov <amonakov@...ras.ru>
To: Fabian Rast <fabian.rast@....de>
cc: musl@...ts.openwall.com
Subject: Re: [PATCH v2] ldso: skip gnu hash calculation if
precomputed
Hello.
Impressive work!
> The other benchmarks were executed in an alpine docker container
If you're using hyperfine just because poop is not packaged, may I suggest
measuring with perf instead? That way, you get cycles/instructions counters
which are invariant w.r.t CPU frequency scaling, so more relevant for us.
'perf stat -r 99 <workload>' runs the workload 99 times and prints deviations.
Alexander
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
