Feedback-ID: ie0a040ee:Fastmail
Message-ID: <83b0c9ce-06ad-4aec-915f-d97ccc0362e3@eisentraut.org>
Date: Tue, 1 Apr 2025 15:30:48 +0200
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: pgsql: Allow parallel CREATE INDEX for GIN indexes
From: Peter Eisentraut <peter@eisentraut.org>
To: Tomas Vondra <tomas.vondra@postgresql.org>,
 pgsql-committers@lists.postgresql.org
References: <E1tp8NA-000bWR-1F@gemulon.postgresql.org>
 <35b1a8bd-5501-407f-b100-22f91834fcc8@eisentraut.org>
Content-Language: en-US
In-Reply-To: <35b1a8bd-5501-407f-b100-22f91834fcc8@eisentraut.org>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: 
 <https://www.postgresql.org/message-id/83b0c9ce-06ad-4aec-915f-d97ccc0362e3%40eisentraut.org>
Precedence: bulk

On 07.03.25 22:22, Peter Eisentraut wrote:
> The new tuplesort_getgintuple() in tuplesortvariants.c has a branch that 
> does "return false" even though the function's return type is GinTuple 
> *.  That is probably a mistake.  Check please.
> 
> Also, this code contains a "pgrminclude ignore", but we don't use those 
> anymore.

Fixed committed.


> On 03.03.25 17:10, Tomas Vondra wrote:
>> Allow parallel CREATE INDEX for GIN indexes
>>
>> Allow using parallel workers to build a GIN index, similarly to BTREE
>> and BRIN. For large tables this may result in significant speedup when
>> the build is CPU-bound.
>>
>> The work is divided so that each worker builds index entries on a subset
>> of the table, determined by the regular parallel scan used to read the
>> data. Each worker uses a local tuplesort to sort and merge the entries
>> for the same key. The TID lists do not overlap (for a given key), which
>> means the merge sort simply concatenates the two lists. The merged
>> entries are written into a shared tuplesort for the leader.
>>
>> The leader needs to merge the sorted entries again, before writing them
>> into the index. But this way a significant part of the work happens in
>> the workers, and the leader is left with merging fewer large entries,
>> which is more efficient.
>>
>> Most of the parallelism infrastructure is a simplified copy of the code
>> used by BTREE indexes, omitting the parts irrelevant for GIN indexes
>> (e.g. uniqueness checks).
>>
>> Original patch by me, with reviews and substantial improvements by
>> Matthias van de Meent, certainly enough to make him a co-author.
>>
>> Author: Tomas Vondra, Matthias van de Meent
>> Reviewed-by: Matthias van de Meent, Andy Fan, Kirill Reshke
>> Discussion: https://postgr.es/m/6ab4003f-a8b8-4d75-a67f- 
>> f25ad98582dc%40enterprisedb.com
>>
>> Branch
>> ------
>> master
>>
>> Details
>> -------
>> https://git.postgresql.org/pg/ 
>> commitdiff/8492feb98f6df3f0f03e84ed56f0d1cbb2ac514c
>>
>> Modified Files
>> --------------
>> src/backend/access/gin/gininsert.c         | 1649 ++++++++++++++++++++ 
>> +++++++-
>> src/backend/access/gin/ginutil.c           |   30 +-
>> src/backend/access/transam/parallel.c      |    4 +
>> src/backend/utils/sort/tuplesortvariants.c |  198 ++++
>> src/include/access/gin.h                   |   15 +
>> src/include/access/gin_private.h           |    1 +
>> src/include/access/gin_tuple.h             |   44 +
>> src/include/utils/tuplesort.h              |    8 +
>> src/tools/pgindent/typedefs.list           |    4 +
>> 9 files changed, 1937 insertions(+), 16 deletions(-)
>>
> 
> 
>