Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:84257
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain zend.com designates 74.125.82.42 as permitted sender)
Content-Type: multipart/alternative;
	boundary=Apple-Mail-62968430-317F-4CD2-BB1B-DDFD83C93EEA
Mime-Version: 1.0 (1.0)
In-Reply-To: <CAAyV7nHF7DfYk58j7EDzjrHZcvBKXizyE4k_NyvF7y_nLVyzwQ@mail.gmail.com>
Date: Tue, 3 Mar 2015 19:48:42 +0200
Cc: PHP Internals <internals@lists.php.net>
Content-Transfer-Encoding: 7bit
Message-ID: <C5DF0613-270C-42CD-B172-75EE3226A2B0@zend.com>
References: <CA+9eiLv+HzuLrECzV37Q1mnNCGNbUj1OocCH3Q7b7PSyWBPtfg@mail.gmail.com>
 <CAAyV7nF53g0wt9Gva5MwULvHDxrqNhLQJMsNTiLP2m0wOAHJkg@mail.gmail.com>
 <ABB24B2E-9456-485E-8510-941076AAAD77@zend.com> <54F08FF3.3040404@seld.be>
 <CA+9eiLumXUD6SowadcfcbtzKYT96ut9fFt-fy5afrGOU1rZhSA@mail.gmail.com>
 <CAAyV7nFF_F+k-9cWqSrLp1vUY+GWxqKqfH6ABZZovZj9d4o8jw@mail.gmail.com>
 <CA+9eiLvZW-myYa9TW6Q0rhLbFRQNLq3iH16FHVyzeThdCUgvGA@mail.gmail.com>
 <63262a9c0edd51bbf38df2a00c87340e@mail.gmail.com> <CAAyV7nHpf8JT0nAF-M9d-65JQHDbYnK7F+c9CUtoyLkkfRiROg@mail.gmail.com>
 <CA+9eiLsa6UpK_e_tRESS6MvXYpbJcFo+Oqd0CC4gByPVNzXNJA@mail.gmail.com>
 <CAAyV7nFvgMa4N9QADMNxBwXtDJMiW=w8RTuTG3zZet6JdmtSSg@mail.gmail.com>
 <CAAyV7nFZGP51J-HHakuHQG0aUK+mS1xQinV8N8MBXE-jqu0_aQ@mail.gmail.com>
 <9977a20c9d756489f41e666d23c89e3f@mail.gmail.com> <CAAyV7nGaf_TC9gMvzr_jDveLKOcMHdKfW1PcJXuo8JxJ32i1DA@mail.gmail.com>
 <9656140ae786d42e7b0da11dbd416a61@mail.gmail.com> <CAAyV7nERsqLgae2hW5k2QfG0cq-XG_5tO9YmE+ejQLrsK7x+DA@mail.gmail.com>
 <CAAyV7nGBDxmbMX=2E4j8s=3NRGcs66w7a2bqiDftAi=wJCna0A@mail.gmail.com>
 <73db72094f98cba6adddfda31e43ab4a@mail.gmail.com> <CAAyV7nHF7DfYk58j7EDzjrHZcvBKXizyE4k_NyvF7y_nLVyzwQ@mail.gmail.com>
To: Anthony Ferrara <ircmaxell@gmail.com>
Subject: Re: [PHP-DEV] Re: Zend JIT Open Sourced
From: zeev@zend.com (Zeev Suraski)

--Apple-Mail-62968430-317F-4CD2-BB1B-DDFD83C93EEA
Content-Type: text/plain;
	charset=utf-8
Content-Transfer-Encoding: quoted-printable


> On 3 =D7=91=D7=9E=D7=A8=D7=A5 2015, at 19:08, Anthony Ferrara <ircmaxell@g=
mail.com> wrote:
>=20
> Zeev,
>=20
> On Tue, Mar 3, 2015 at 11:32 AM, Zeev Suraski <zeev@zend.com> wrote:
>>> It wins on uneven ground.
>>=20
>> Without the output buffering code it's a LOT more uneven, as the PHP vers=
ion
>> is flushing every byte - approx. 6,000 such flushes.
>=20
> Then the benchmark should suffer because of that. You're testing the
> full stack of all of the other languages, yet disabling a significant
> part of it for PHP and saying "it's still fair because that part of
> the stack is expensive".

Yes, the part that has nothing at all to do with the benchmark at hand, and t=
hat changes on a per-SAPI basis and creates nothing but unrelated noise for t=
his benchmark.  As I repeatedly said, and as should be obvious to anybody lo=
oking at cpu-bound benchmarks like Mandelbrot, the idea is to focus on the a=
lgorithm and factor out everything else, which is exactly what the code for b=
oth PHP and C (as well as the other langs) does.  If you haven't yet - run t=
he new line-free C version that has full buffering.  Spoiler: it has identic=
al performance.


Zeev=

--Apple-Mail-62968430-317F-4CD2-BB1B-DDFD83C93EEA--