[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <0d40e6f8-9673-30d8-851f-4a47874fa34e@ispras.ru>
Date: Mon, 22 Dec 2025 21:40:12 +0300 (MSK)
From: Alexander Monakov <amonakov@...ras.ru>
To: Fabian Rast <fabian.rast@....de>
cc: Rich Felker <dalias@...c.org>, musl@...ts.openwall.com
Subject: Re: [PATCH v2] ldso: skip gnu hash calculation if
precomputed
On Mon, 22 Dec 2025, Fabian Rast wrote:
> /usr/lib/libwebkit2gtk-4.1.so.0
> cycles: 234214481 (0.16) -> 224880118 (0.11) -3.99%
> instructions: 329852233 (0.0) -> 233217195 (0.0) -29.3%
> ref-cycles: 120347198 (1.09) -> 135856350 (0.6) 12.89%
> duration_time: 62053249 (1.08) -> 69948475 (0.59) 12.72%
>
[...]
>
> Interestingly, there is a big regression in perfomance loading libwebkit2gtk
> that i cannot explain...
You have a reduction in actual CPU cycles (-4%, similar to others) and an
increase of the reference counter cycles (and the matching increase in
wall-clock time), which means that your CPU was running at a lower frequency
in this test. If it is reproducible, perhaps it's an artifact of how CPU's
automatic frequency scaling works: without the hashing loop, most of the
remaining work is low-IPC (lots of TLB and cache misses), so maybe it doesn't
raise clocks in this case. You can test e.g. by running at a fixed frequency.
(you may already know, but just in case: cpupower tool can be used for
requesting a fixed frequency with the 'performance' governor; to disable
turbo clocks on AMD CPUs, write '0' to /sys/devices/system/cpu/cpufreq/boost )
Thank you for your continued work on this!
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.
