MIME-Version: 1.0
References: <ZyMg8zXEfJZ74ZBX@paquier.xyz> <CAApHDvrXzPAr3FxoBuB7b3D-okNoNA2jxLun1rW8Yw5wkbqusw@mail.gmail.com>
 <ZyR2Tzdmj9c6E/QB@ip-10-97-1-34.eu-west-3.compute.internal>
 <CAApHDvrEDWPEyWQF7UhAt+8JR6w2eVpv1xP21FqFH+c=8Lb3JA@mail.gmail.com>
 <ZySFDUu5-1jConZE@paquier.xyz> <ZySG1uHmmgQT+dpE@ip-10-97-1-34.eu-west-3.compute.internal>
 <ZySIAShKvyGxdc8n@paquier.xyz> <CAApHDvqULM7KhreSa8PU2bshrguOaRdJsfjKCTvDkByYRFLaMg@mail.gmail.com>
 <ZyTOgpIlQTgMBrah@ip-10-97-1-34.eu-west-3.compute.internal>
 <CAApHDvoU1SfD6rWjGp2tpLd+VGg_TfXhPJgCSNY5Z5LULXiv1A@mail.gmail.com> <ZykBwloj77DVlhnN@ip-10-97-1-34.eu-west-3.compute.internal>
In-Reply-To: <ZykBwloj77DVlhnN@ip-10-97-1-34.eu-west-3.compute.internal>
From: Ranier Vilela <ranier.vf@gmail.com>
Date: Mon, 4 Nov 2024 14:38:54 -0300
Message-ID: <CAEudQAqteX_B_2a+qRfG3gzOP5vOGTmihHkXCbC_FhjuZhfGhw@mail.gmail.com>
Subject: Re: define pg_structiszero(addr, s, r)
To: Bertrand Drouvot <bertranddrouvot.pg@gmail.com>
Cc: David Rowley <dgrowleyml@gmail.com>, Michael Paquier <michael@paquier.xyz>, 
	Peter Smith <smithpb2250@gmail.com>, Peter Eisentraut <peter@eisentraut.org>, 
	Heikki Linnakangas <hlinnaka@iki.fi>, pgsql-hackers@lists.postgresql.org
Content-Type: multipart/alternative; boundary="000000000000eb8b2a062619c29f"
Archived-At: <https://www.postgresql.org/message-id/CAEudQAqteX_B_2a%2BqRfG3gzOP5vOGTmihHkXCbC_FhjuZhfGhw%40mail.gmail.com>
Precedence: bulk

--000000000000eb8b2a062619c29f
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Hi.

Em seg., 4 de nov. de 2024 =C3=A0s 14:18, Bertrand Drouvot <
bertranddrouvot.pg@gmail.com> escreveu:

> Hi,
>
> On Tue, Nov 05, 2024 at 12:24:48AM +1300, David Rowley wrote:
> > On Sat, 2 Nov 2024 at 01:50, Bertrand Drouvot
> > <bertranddrouvot.pg@gmail.com> wrote:
> > >
> > > On Fri, Nov 01, 2024 at 09:47:05PM +1300, David Rowley wrote:
> > > > I've attached what I thought a more optimal version might look like
> in
> > > > case anyone thinks making it better is a good idea.
> > > >
> > >
> > > Thanks for the proposal!
> > >
> > > I like the idea, I think that's worth to add a few comments, somethin=
g
> like:
> >
> > I'm happy if you want to pick this up and continue working on it.
>
> Sure, please find attached v1, the changes are:
>
> - switch from "const char" to "const unsigned char" (could have been done
> in the
> current version of pg_memory_is_all_zeros() though)
> - added some comments
> - adding an Assert for ptr !=3D 0
> - re-introduce the function call in PageIsVerifiedExtended()
> - propose a commit message
>

I think we can add a small optimization to this last patch [1].
The variable *aligned_end* is only needed in the second loop (for).
So, only before the for loop do we actually declare it.

Result before this change:
check zeros using BERTRAND 1 0.000031s

Result after this change:
check zeros using BERTRAND 1 0.000018s

+ const unsigned char *aligned_end;

+ /* Multiple bytes comparison(s) at once */
+ aligned_end =3D (const unsigned char *) ((uintptr_t) end &
(~(sizeof(size_t) - 1)));
+ for (; p < aligned_end; p +=3D sizeof(size_t))

best regards,
Ranier Vilela

[1] https://godbolt.org/z/fvdzjbnc7

--000000000000eb8b2a062619c29f
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>Hi.<br></div><br><div class=3D"gmail_quote"><div dir=
=3D"ltr" class=3D"gmail_attr">Em seg., 4 de nov. de 2024 =C3=A0s 14:18, Ber=
trand Drouvot &lt;<a href=3D"mailto:bertranddrouvot.pg@gmail.com">bertrandd=
rouvot.pg@gmail.com</a>&gt; escreveu:<br></div><blockquote class=3D"gmail_q=
uote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,2=
04);padding-left:1ex">Hi,<br>
<br>
On Tue, Nov 05, 2024 at 12:24:48AM +1300, David Rowley wrote:<br>
&gt; On Sat, 2 Nov 2024 at 01:50, Bertrand Drouvot<br>
&gt; &lt;<a href=3D"mailto:bertranddrouvot.pg@gmail.com" target=3D"_blank">=
bertranddrouvot.pg@gmail.com</a>&gt; wrote:<br>
&gt; &gt;<br>
&gt; &gt; On Fri, Nov 01, 2024 at 09:47:05PM +1300, David Rowley wrote:<br>
&gt; &gt; &gt; I&#39;ve attached what I thought a more optimal version migh=
t look like in<br>
&gt; &gt; &gt; case anyone thinks making it better is a good idea.<br>
&gt; &gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; Thanks for the proposal!<br>
&gt; &gt;<br>
&gt; &gt; I like the idea, I think that&#39;s worth to add a few comments, =
something like:<br>
&gt; <br>
&gt; I&#39;m happy if you want to pick this up and continue working on it.<=
br>
<br>
Sure, please find attached v1, the changes are:<br>
<br>
- switch from &quot;const char&quot; to &quot;const unsigned char&quot; (co=
uld have been done in the<br>
current version of pg_memory_is_all_zeros() though)<br>
- added some comments<br>
- adding an Assert for ptr !=3D 0<br>
- re-introduce the function call in PageIsVerifiedExtended()<br>
- propose a commit message<br></blockquote><div><br></div><div>I think we c=
an add a small optimization to this last patch [1].<br>The variable *aligne=
d_end* is only needed in the second loop (for).<br>So, only before the for =
loop do we actually declare it.<br><br>Result before this change:<br>check =
zeros using BERTRAND 1 0.000031s<br><br>Result after this change:<br>check =
zeros using BERTRAND 1 0.000018s</div><div><br></div><div>+	const unsigned =
char *aligned_end;</div><div><br></div><div>+	/* Multiple bytes comparison(=
s) at once */<br>+	aligned_end =3D (const unsigned char *) ((uintptr_t) end=
 &amp; (~(sizeof(size_t) - 1)));<br>+	for (; p &lt; aligned_end; p +=3D siz=
eof(size_t))</div><div><br></div></div><div class=3D"gmail_quote">best rega=
rds,</div><div class=3D"gmail_quote">Ranier Vilela</div><div class=3D"gmail=
_quote"><br></div><div class=3D"gmail_quote">[1] <a href=3D"https://godbolt=
.org/z/fvdzjbnc7">https://godbolt.org/z/fvdzjbnc7</a></div></div>

--000000000000eb8b2a062619c29f--