Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:115435
MIME-Version: 1.0
References: <CAFv4g+HRHrvCOTgpbVfCXj06De1kL+1+RwnGgq6tcyLNAtLfQA@mail.gmail.com>
 <CAF+90c8DbBajg7yukYkyXeJp2RU3JngLcr_zJdP=tOw9ebb6ag@mail.gmail.com> <CAH5C8xUzQiRm8SGYN2kuQHefyO-s+DfQY=2PktMMBRVqoq96jg@mail.gmail.com>
In-Reply-To: <CAH5C8xUzQiRm8SGYN2kuQHefyO-s+DfQY=2PktMMBRVqoq96jg@mail.gmail.com>
Date: Fri, 16 Jul 2021 12:04:04 +0200
Message-ID: <CAF+90c_ft=HgZAu-F5n6awuizZs1XX4N50LhPiK17fktSgedfA@mail.gmail.com>
To: =?UTF-8?B?TcOhdMOpIEtvY3Npcw==?= <kocsismate90@gmail.com>
Cc: Craig Francis <craig@craigfrancis.co.uk>, PHP internals <internals@lists.php.net>
Content-Type: multipart/alternative; boundary="0000000000001a323305c73ab66f"
Subject: Re: [PHP-DEV] [RFC] [VOTE] is_literal
From: nikita.ppv@gmail.com (Nikita Popov)

--0000000000001a323305c73ab66f
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Fri, Jul 9, 2021 at 12:23 AM M=C3=A1t=C3=A9 Kocsis <kocsismate90@gmail.c=
om> wrote:

> Hi Nikita,
>
> I performed a few other benchmarks in order to provide a little bit more
> insights into the performance aspect of the RFC. My latest measurement is
> using the same setup as the previous
> one, although I made a few smaller changes in the process (e.g. changing
> the order of the tests). So at the end, I got the following results:
>
> - Laravel demo app: +0.1%
> - Symfony demo app: +0.43%
> - bench.php: +0.4%
> - custom concat bench: +3.89%
>
> The results are comparing the "literals" branch with the commit in master
> on which Joe's work is based. As far as I remember, the simpler variant o=
f
> is_literals had a 0.1% slowdown
> in case of Symfony, and 2.09% in case of the custom concat benchmark.
> Personally, I think even the current numbers are acceptable, especially
> because PHP 8.1 is going to be
> roughly 30% faster than PHP 8.0 according to my benchmark in case of the
> Symfony demo app. Laravel's result is very similar, but I don't remember
> about the exact numbers.
>

Thanks M=C3=A1t=C3=A9! While I'm not really happy with the technical implic=
ations,
your numbers clearly show that there is barely any impact in practice. I've
dropped my "no" vote for that reason.

Regards,
Nikita

>

--0000000000001a323305c73ab66f--