<div dir="auto">That's really a fantastic news! Thanks Han again for your work on this, and thanks to all the people that helped with this!<div dir="auto"><br></div><div dir="auto">Giuseppe. </div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Sun, 28 Nov 2021, 20:26 Regina Obe, <<a href="mailto:lr@pcorp.us">lr@pcorp.us</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">I think Han and Kontur have some benchmarks.<br>
<br>
<a href="https://git.osgeo.org/gitea/postgis/postgis-performance" rel="noreferrer noreferrer" target="_blank">https://git.osgeo.org/gitea/postgis/postgis-performance</a> <br>
<br>
In my pgTap benchmarks the index build timing is approximately 3-4 times<br>
faster with osm data of Belarus and China. Han said the query performance<br>
slowed a little, but I didn't see that in my pgTap tests (no really<br>
noticeable difference with intersect checking), so might have been fixed by<br>
Kontur in their cleanup before I ran these tests or I'm missing a test.<br>
<br>
Here are the pgTap benchmarks based on alpha1 with PG 14.0rc1, nothing with<br>
the index has changed since so should be representative. <br>
<br>
<a href="https://git.osgeo.org/gitea/postgis/postgis-performance/src/branch/master/pgtap/postgis/03_build_gist_index.sql" rel="noreferrer noreferrer" target="_blank">https://git.osgeo.org/gitea/postgis/postgis-performance/src/branch/master/pg<br>
tap/postgis/03_build_gist_index.sql</a> <br>
-- index building compare --<br>
PostGIS 3.1.4<br>
[08:45:51] pgtap/postgis/03_build_gist_index.sql .... <br>
1..3<br>
ok 1 - Index built: multipolygons<br>
ok 2 - Index built: lines<br>
ok 3 - Index built: points<br>
ok 70742 ms ( 0.01 usr 0.00 sys + 0.00 cusr 0.01 csys = 0.02 CPU)<br>
<br>
<br>
PostGIS 3.2.0alpha1<br>
[09:17:54] pgtap/postgis/03_build_gist_index.sql .... <br>
1..3<br>
ok 1 - Index built: multipolygons<br>
ok 2 - Index built: lines<br>
ok 3 - Index built: points<br>
ok 17029 ms ( 0.00 usr 0.00 sys + 0.00 cusr 0.01 csys = 0.01 CPU)<br>
<br>
<br>
-- intersect checks compare<br>
<a href="https://git.osgeo.org/gitea/postgis/postgis-performance/src/branch/master/pgtap/postgis/04_index_intersect.sql" rel="noreferrer noreferrer" target="_blank">https://git.osgeo.org/gitea/postgis/postgis-performance/src/branch/master/pg<br>
tap/postgis/04_index_intersect.sql</a><br>
<br>
-- POSTGIS "3.1.4"<br>
[08:46:16] pgtap/postgis/04_index_intersect.sql ..... <br>
1..9<br>
ok 1 - Check has osm_belarus.multipolygons has gist<br>
ok 2 - Check has osm_belarus.lines geom gist<br>
ok 3 - Check has osm_belarus.points geom gist<br>
ok 4 - intersect multipolygon check no match with index<br>
ok 5 - intersect multipolygon check match with index<br>
ok 6 - intersect linestring check no match with index<br>
ok 7 - intersect linestring check match with index<br>
ok 8 - intersect pointcheck no match with index<br>
ok 9 - intersect point check match with index<br>
ok 24981 ms ( 0.00 usr 0.00 sys + 0.02 cusr 0.00 csys = 0.02 CPU)<br>
<br>
-- POSTGIS="3.2.0alpha1 3.2.0alpha1"<br>
[09:18:16] pgtap/postgis/04_index_intersect.sql ..... <br>
1..9<br>
ok 1 - Check has osm_belarus.multipolygons has gist<br>
ok 2 - Check has osm_belarus.lines geom gist<br>
ok 3 - Check has osm_belarus.points geom gist<br>
ok 4 - intersect multipolygon check no match with index<br>
ok 5 - intersect multipolygon check match with index<br>
ok 6 - intersect linestring check no match with index<br>
ok 7 - intersect linestring check match with index<br>
ok 8 - intersect pointcheck no match with index<br>
ok 9 - intersect point check match with index<br>
ok 22459 ms ( 0.00 usr 0.00 sys + 0.00 cusr 0.01 csys = 0.01 CPU)<br>
<br>
> -----Original Message-----<br>
> From: postgis-users [mailto:<a href="mailto:postgis-users-bounces@lists.osgeo.org" target="_blank" rel="noreferrer">postgis-users-bounces@lists.osgeo.org</a>] On<br>
Behalf<br>
> Of Marco Boeringa<br>
> Sent: Saturday, November 27, 2021 4:27 PM<br>
> To: PostGIS Users Discussion <<a href="mailto:postgis-users@lists.osgeo.org" target="_blank" rel="noreferrer">postgis-users@lists.osgeo.org</a>><br>
> Subject: [postgis-users] Experiences with the new faster GiST building in<br>
> PostGIS 3.2 and PostgreSQL 14?<br>
> <br>
> Hi all,<br>
> <br>
> Now with the betas out for PostGIS 3.2, I was wondering if anyone has some<br>
> real world comparative figures to share for creating spatial indexes using<br>
GiST<br>
> in PostGIS 3.2 and PostgreSQL 14, versus the older GiST implementation in<br>
> previous versions?<br>
> <br>
> Creating spatial indexes on e.g. OpenStreetMap sized datasets is still an<br>
> expensive operation, and any gains in the latest releases would be highly<br>
> welcome. Has someone done a comparison between releases that might give<br>
> all of us that haven't yet upgraded, an idea of possible benefits of the<br>
new<br>
> implementation?<br>
> <br>
> Marco<br>
> <br>
> _______________________________________________<br>
> postgis-users mailing list<br>
> <a href="mailto:postgis-users@lists.osgeo.org" target="_blank" rel="noreferrer">postgis-users@lists.osgeo.org</a><br>
> <a href="https://lists.osgeo.org/mailman/listinfo/postgis-users" rel="noreferrer noreferrer" target="_blank">https://lists.osgeo.org/mailman/listinfo/postgis-users</a><br>
<br>
_______________________________________________<br>
postgis-users mailing list<br>
<a href="mailto:postgis-users@lists.osgeo.org" target="_blank" rel="noreferrer">postgis-users@lists.osgeo.org</a><br>
<a href="https://lists.osgeo.org/mailman/listinfo/postgis-users" rel="noreferrer noreferrer" target="_blank">https://lists.osgeo.org/mailman/listinfo/postgis-users</a><br>
</blockquote></div>