Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:127453
Feedback-ID: ifab94697:Fastmail
Precedence: bulk
MIME-Version: 1.0
Date: Sun, 25 May 2025 15:28:18 +0200
To: "Alwin Garside" <alwin@garsi.de>,
 "Daniel Kesselberg" <mail@danielkesselberg.de>
Cc: "PHP Internals" <internals@lists.php.net>
Message-ID: <fa495912-6669-449d-b573-fea7b29075cd@app.fastmail.com>
In-Reply-To: <60B4033B-2E6A-4157-A698-6C58454B07C9@garsi.de>
References: <bc1ec15125844c78a6aa23dce6148af9@danielkesselberg.de>
 <a1509633-ad50-4779-bc46-0f782cd8b7c5@app.fastmail.com>
 <60B4033B-2E6A-4157-A698-6C58454B07C9@garsi.de>
Subject: Re: [PHP-DEV] [RFC] Add num_available_processors
Content-Type: multipart/alternative;
 boundary=bbfb6805193c4d4b8b477a87d5de5cd4
From: rob@bottled.codes ("Rob Landers")

--bbfb6805193c4d4b8b477a87d5de5cd4
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

On Sun, May 25, 2025, at 12:07, Alwin Garside wrote:
> On 24 May 2025, at 20:48, Rob Landers <rob@bottled.codes> wrote:
> >=20
> > On Sat, May 24, 2025, at 19:37, Daniel Kesselberg wrote:
> >>=20
> >>=20
> >> Hi everyone,
> >>=20
> >> I'm happy to share my first RFC :) It proposes adding a small funct=
ion=20
> >> to retrieve the number of available processors; a feature that's=20
> >> commonly found in other programming languages and one that I believ=
e=20
> >> would be a useful addition to PHP.
> >>=20
> >> The related PR has already received a bit of early traction, and no=
w=20
> >> that the RFC is complete, I'm looking forward to your feedback!
> >>=20
> >> RFC: https://wiki.php.net/RFC/num_available_processors
> >> Patch: https://github.com/php/php-src/pull/11137
> >>=20
> >> Best
> >> Daniel
> >>=20
> >=20
> > Looks good!
> >=20
> > My main question is: what is this actually counting? In the RFC it m=
entions "available processing units" ... which means, what? What counts =
as a "processing unit"? Are we talking about CPU Threads/cores; NPU core=
s; TPM cores; clocks? GPS? GPU? ... a modern computer has many "processi=
ng units" for different purposes and workloads. I=E2=80=99m assuming thi=
s is CPU Threads, not physical cores? I will refer to CPU Threads as "Lo=
gical Cores" so we all don=E2=80=99t get confused since most of us here =
are programmers and saying "thread" has a different meaning.
> >=20
> > Secondly, how is it counting "available"? If I assign PHP to a speci=
fic CPU affinity mask (say one logical core), will it return 1, or the t=
otal number of logical cores available on my machine? I would expect it =
to be 1, since PHP only has access to 1, but I can also see the logic in=
 returning the total number.
> >=20
> > =E2=80=94 Rob
>=20
>=20
> Hi Daniel,
>=20
> I agree with Rob that "processor" is a bit too ambiguous. I'd use the =
phrase "cpu_core" instead. Yes, technically that's not entirely accurate=
 when hyper-threading is used, but in most cases it's not trivial to dis=
tinguish physical cores from logical cores anyway, and "cpu_cores" provi=
des the most understandable abstraction for the vast majority of use cas=
es: deciding how many parallel processes one should use for optimal use =
of the CPU.

Yes, that is why I was curious as to the method it was using to return t=
he number. Using the sysconf method will just return the total number of=
 available cores. However, in Linux there is also the sched_getaffinity =
method which returns the total number of cores allocated to the current =
process (which may be lower). For example, when sharing a webserver and =
a worker queue on a single system, I may only allocate 1/3 of my cores (=
via `taskset`) to the worker queue, and 1/3 cores to the webserver and t=
he last 1/3 for other tasks (such as a database). This prevents the work=
er queue from taking over the entire system and slowing down my web requ=
ests and database.

If we have 12 cores and I=E2=80=99ve only allocated 4 cores for PHP, I w=
ould want this function to tell my worker queue that it only has 4 cores=
, using the system above =E2=80=94 not 12. This probably doesn=E2=80=99t=
 matter too much when this is the only process on the machine. However, =
in any serious production environment, you want to be able to prevent a =
run-away process/job from harming other unrelated aspects of the system.

You can also see this behaviour with nproc:

=E2=9D=AF taskset -c 0 nproc
1

=E2=9D=AF nproc
16



=E2=80=94 Rob
--bbfb6805193c4d4b8b477a87d5de5cd4
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE html><html><head><title></title></head><body><div>On Sun, May =
25, 2025, at 12:07, Alwin Garside wrote:</div><blockquote type=3D"cite" =
id=3D"qt" style=3D""><div>On 24 May 2025, at 20:48, Rob Landers &lt;<a h=
ref=3D"mailto:rob@bottled.codes">rob@bottled.codes</a>&gt; wrote:</div><=
div>&gt;&nbsp;</div><div>&gt; On Sat, May 24, 2025, at 19:37, Daniel Kes=
selberg wrote:</div><div>&gt;&gt;&nbsp;</div><div>&gt;&gt;&nbsp;</div><d=
iv>&gt;&gt; Hi everyone,</div><div>&gt;&gt;&nbsp;</div><div>&gt;&gt; I'm=
 happy to share my first RFC :) It proposes adding a small function&nbsp=
;</div><div>&gt;&gt; to retrieve the number of available processors; a f=
eature that's&nbsp;</div><div>&gt;&gt; commonly found in other programmi=
ng languages and one that I believe&nbsp;</div><div>&gt;&gt; would be a =
useful addition to PHP.</div><div>&gt;&gt;&nbsp;</div><div>&gt;&gt; The =
related PR has already received a bit of early traction, and now&nbsp;</=
div><div>&gt;&gt; that the RFC is complete, I'm looking forward to your =
feedback!</div><div>&gt;&gt;&nbsp;</div><div>&gt;&gt; RFC:&nbsp;<a href=3D=
"https://wiki.php.net/RFC/num_available_processors">https://wiki.php.net=
/RFC/num_available_processors</a></div><div>&gt;&gt; Patch:&nbsp;<a href=
=3D"https://github.com/php/php-src/pull/11137">https://github.com/php/ph=
p-src/pull/11137</a></div><div>&gt;&gt;&nbsp;</div><div>&gt;&gt; Best</d=
iv><div>&gt;&gt; Daniel</div><div>&gt;&gt;&nbsp;</div><div>&gt;&nbsp;</d=
iv><div>&gt; Looks good!</div><div>&gt;&nbsp;</div><div>&gt; My main que=
stion is: what is this actually counting? In the RFC it mentions "availa=
ble processing units" ... which means, what? What counts as a "processin=
g unit"? Are we talking about CPU Threads/cores; NPU cores; TPM cores; c=
locks? GPS? GPU? ... a modern computer has many "processing units" for d=
ifferent purposes and workloads. I=E2=80=99m assuming this is CPU Thread=
s, not physical cores? I will refer to CPU Threads as "Logical Cores" so=
 we all don=E2=80=99t get confused since most of us here are programmers=
 and saying "thread" has a different meaning.</div><div>&gt;&nbsp;</div>=
<div>&gt; Secondly, how is it counting "available"? If I assign PHP to a=
 specific CPU affinity mask (say one logical core), will it return 1, or=
 the total number of logical cores available on my machine? I would expe=
ct it to be 1, since PHP only has access to 1, but I can also see the lo=
gic in returning the total number.</div><div>&gt;&nbsp;</div><div>&gt; =E2=
=80=94 Rob</div><div><br></div><div><br></div><div>Hi Daniel,</div><div>=
<br></div><div>I agree with Rob that "processor" is a bit too ambiguous.=
 I'd use the phrase "cpu_core" instead. Yes, technically that's not enti=
rely accurate when hyper-threading is used, but in most cases it's not t=
rivial to distinguish physical cores from logical cores anyway, and "cpu=
_cores" provides the most understandable abstraction for the vast majori=
ty of use cases: deciding how many parallel processes one should use for=
 optimal use of the CPU.</div></blockquote><div><br></div><div>Yes, that=
 is why I was curious as to the method it was using to return the number=
. Using the sysconf method will just return the total number of availabl=
e cores. However, in Linux there is also the&nbsp;sched_getaffinity meth=
od which returns the total number of cores allocated to the current proc=
ess (which may be lower). For example, when sharing a webserver and a wo=
rker queue on a single system, I may only allocate 1/3 of my cores (via =
`taskset`) to the worker queue, and 1/3 cores to the webserver and the l=
ast 1/3 for other tasks (such as a database). This prevents the worker q=
ueue from taking over the entire system and slowing down my web requests=
 and database.</div><div><br></div><div>If we have 12 cores and I=E2=80=99=
ve only allocated 4 cores for PHP, I would want this function to tell my=
 worker queue that it only has 4 cores, using the system above =E2=80=94=
 not 12. This probably doesn=E2=80=99t matter too much when this is the =
only process on the machine. However, in any serious production environm=
ent, you want to be able to prevent a run-away process/job from harming =
other unrelated aspects of the system.</div><div><br></div><div>You can =
also see this behaviour with nproc:</div><div><br></div><div>=E2=9D=AF t=
askset -c 0 nproc</div><div>1</div><div><br></div><div>=E2=9D=AF nproc</=
div><div>16</div><div><br></div><div><br></div><div><br></div><div id=3D=
"sig121229152">=E2=80=94 Rob</div></body></html>
--bbfb6805193c4d4b8b477a87d5de5cd4--