Date: Mon, 15 Aug 2022 14:23:04 -0700
From: Nathan Bossart <nathandbossart@gmail.com>
To: John Naylor <john.naylor@enterprisedb.com>
Cc: Andres Freund <andres@anarazel.de>, Tom Lane <tgl@sss.pgh.pa.us>,
	Jelte Fennema <Jelte.Fennema@microsoft.com>,
	"pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>,
	Merlin Moncure <mmoncure@gmail.com>,
	Andrew Dunstan <andrew.dunstan@2ndquadrant.com>,
	Stephen Frost <sfrost@snowman.net>
Subject: Re: [PATCH] Optimize json_lex_string by batching character copying
Message-ID: <20220815212304.GA251709@nathanxps13>
References: <20220625001810.q4giojmraweoyl5e@alap3.anarazel.de>
 <20220625010551.r5qrk5n2euuwqos3@alap3.anarazel.de>
 <CAFBsxsGhaR2KQ5eisaK=6Vm60t=axhD8Ckj1qFoCH1pktZi+2w@mail.gmail.com>
 <20220706051801.gfbzjxgflbsglyt6@awork3.anarazel.de>
 <CAFBsxsE0FkMT0vZOLfJEOCM1Di81Wy_5fE9S1rn0PNHMYrmCCw@mail.gmail.com>
 <20220711154150.hwzw7qw2k6iahofy@awork3.anarazel.de>
 <1119852.1657554806@sss.pgh.pa.us>
 <20220711160728.tljbnowfgxefdbli@awork3.anarazel.de>
 <CAFBsxsGzaaGLF=Nuq61iRXTyspbO9rOjhSqFN=V6ozzmta5mXg@mail.gmail.com>
 <CAFBsxsESLUyJ5spfOSyPrOvKUEYYNqsBosue9SV1j8ecgNXSKA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: 
 <CAFBsxsESLUyJ5spfOSyPrOvKUEYYNqsBosue9SV1j8ecgNXSKA@mail.gmail.com>
Archived-At: 
 <https://www.postgresql.org/message-id/20220815212304.GA251709%40nathanxps13>
Precedence: bulk

On Mon, Aug 15, 2022 at 08:33:21PM +0700, John Naylor wrote:
> The attached implements the above, more or less, using new pg_lfind8()
> and pg_lfind8_le(), which in turn are based on helper functions that
> act on a single vector. The pg_lfind* functions have regression tests,
> but I haven't done the same for json yet. I went the extra step to use
> bit-twiddling for non-SSE builds using uint64 as a "vector", which
> still gives a pretty good boost (test below, min of 3):

Looks pretty reasonable to me.

> +#ifdef USE_SSE2
> +		chunk = _mm_loadu_si128((const __m128i *) &base[i]);
> +#else
> +		memcpy(&chunk, &base[i], sizeof(chunk));
> +#endif							/* USE_SSE2 */

> +#ifdef USE_SSE2
> +		chunk = _mm_loadu_si128((const __m128i *) &base[i]);
> +#else
> +		memcpy(&chunk, &base[i], sizeof(chunk));
> +#endif							/* USE_SSE2 */

Perhaps there should be a macro or inline function for loading a vector so
that these USE_SSE2 checks can be abstracted away, too.

-- 
Nathan Bossart
Amazon Web Services: https://aws.amazon.com