Re: [PATCH v2 08/10] generate-cmdlist.sh: do not shell out to "sed"

From: Jeff King <peff@peff.net>
Date: Oct 25, 2021, 16:46 UTC
Message-ID: <YXbfdEqKHc5b19u9@coredump.intra.peff.net>
In-Reply-To: <patch-v2-08.10-83318d6c0da-20211022T193027Z-avarab@gmail.com>

On Fri, Oct 22, 2021 at 09:36:12PM +0200, Ævar Arnfjörð Bjarmason wrote:

Show 5 quoted lines

> From: Jeff King <peff@peff.net>
> 
> Replace the "sed" invocation in get_synopsis() with a pure-shell
> version. This speeds up generate-cmdlist.sh significantly. Compared to
> HEAD~ (old) and "master" we are, according to hyperfine(1):

Unsurprisingly I'm in favor of this. ;)

Curiously again, I get more dramatic results than you:

>   'sh generate-cmdlist.sh command-list.txt' ran
>    12.69 ± 5.01 times faster than 'sh generate-cmdlist.sh.old command-list.txt'
>    18.34 ± 3.03 times faster than 'sh generate-cmdlist.sh.master command-list.txt'

    'sh generate-cmdlist.sh command-list.txt' ran
     22.44 ± 13.59 times faster than 'sh generate-cmdlist.sh.old command-list.txt'
     57.35 ± 34.10 times faster than 'sh generate-cmdlist.sh.master command-list.txt'

It's like spawning processes is somehow faster on your machine than mine. I wonder if it's a CPU governor thing. This is a laptop, and those numbers come from using "powersave". Doing "cpufreq-set -g performance", I get:

   'sh generate-cmdlist.sh command-list.txt' ran
   14.35 ± 0.23 times faster than 'sh generate-cmdlist.sh.old command-list.txt'
   33.15 ± 0.50 times faster than 'sh generate-cmdlist.sh.master command-list.txt'

which is closer. But most notably all versions are 3-5x faster than their "powersave" counterparts. I wonder if that has been driving some of the confusion in our timings in this thread.

Either way, I think this is still a good direction to go.

-Peff