QGIS performance problem

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

QGIS performance problem

Brent Wood
Hi,

I have a PostGIS table & shapefile of the New Zealand coast (1:50,000)

If I open this table it plots in a fraction of a second, pretty much as expected, with similar speeds for both the shapefile & table.

However, if I zoom in, the redraw takes three minutes or more.

While I'm waiting for the refresh, Qgis is showing 100% cpu load, postgres isn't showing at all (top). On an AMD dual core 3.2Ghz cpu with 8Gb ram it should not take anything like this long, and with earlier versions of Qgis (pre 1.0) it was much faster.

Performance is similar with both table & shapefile data sources, so it does not seem Postgres related.

However, plotting lines, points or polygons with tens of thousands of features rendered works well, it seems to primarily be polygons with many vertices (not even any holes) that are the problem.

OpenSuse 11.1, Qgis 1.2.0, Postgis table described below, only 3 features/records, indexed & vacuumed.


Does anyone have any idea what may be the issue?

I can provide an ssh login if anyone wants to test this for themselves.


Thanks,

   Brent Wood






                                 Table "public.coast_poly"
 Column  |         Type          |                        Modifiers                        
---------+-----------------------+----------------------------------------------------------
 gid     | integer               | not null default nextval('coast_poly_gid_seq'::regclass)
 id      | bigint                |
 fid     | bigint                |
 name    | character varying(42) |
 geom_ll | geometry              |
Indexes:
    "coast_poly_pkey" PRIMARY KEY, btree (gid)
    "coast_poly_geom_ll_gist" gist (geom_ll)
Check constraints:
    "enforce_dims_geom_ll" CHECK (ndims(geom_ll) = 2)
    "enforce_geotype_geom_ll" CHECK (geometrytype(geom_ll) = 'MULTIPOLYGON'::text OR geom_ll IS NULL)
    "enforce_srid_geom_ll" CHECK (srid(geom_ll) = 4326)


_______________________________________________
Qgis-user mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/qgis-user
Reply | Threaded
Open this post in threaded view
|

Re: QGIS performance problem

Andreas Neumann-4
Hi Brent,

I have similar behaviour. It happens if polygons with lots of vertices
are involved. Things get even worse if you try to print this data with
high resolution.

Marco Hugentobler and Jürgen Fischer are already working on this issue
these days. They can also better explain the issues we are running into
- seems to be qt and clipping related when bounding boxes of big
polygons are touched when zooming in or printing on high resolutions.

I am glad to hear that I am not the only person experiencing these
problems - seems to be a real issue.

Andreas

[hidden email] wrote:

> Hi,
>
> I have a PostGIS table & shapefile of the New Zealand coast (1:50,000)
>
> If I open this table it plots in a fraction of a second, pretty much as expected, with similar speeds for both the shapefile & table.
>
> However, if I zoom in, the redraw takes three minutes or more.
>
> While I'm waiting for the refresh, Qgis is showing 100% cpu load, postgres isn't showing at all (top). On an AMD dual core 3.2Ghz cpu with 8Gb ram it should not take anything like this long, and with earlier versions of Qgis (pre 1.0) it was much faster.
>
> Performance is similar with both table & shapefile data sources, so it does not seem Postgres related.
>
> However, plotting lines, points or polygons with tens of thousands of features rendered works well, it seems to primarily be polygons with many vertices (not even any holes) that are the problem.
>
> OpenSuse 11.1, Qgis 1.2.0, Postgis table described below, only 3 features/records, indexed & vacuumed.
>
>
> Does anyone have any idea what may be the issue?
>
> I can provide an ssh login if anyone wants to test this for themselves.
>
>
> Thanks,
>
>    Brent Wood
>
>
>
>
>
>
>                                  Table "public.coast_poly"
>  Column  |         Type          |                        Modifiers                        
> ---------+-----------------------+----------------------------------------------------------
>  gid     | integer               | not null default nextval('coast_poly_gid_seq'::regclass)
>  id      | bigint                |
>  fid     | bigint                |
>  name    | character varying(42) |
>  geom_ll | geometry              |
> Indexes:
>     "coast_poly_pkey" PRIMARY KEY, btree (gid)
>     "coast_poly_geom_ll_gist" gist (geom_ll)
> Check constraints:
>     "enforce_dims_geom_ll" CHECK (ndims(geom_ll) = 2)
>     "enforce_geotype_geom_ll" CHECK (geometrytype(geom_ll) = 'MULTIPOLYGON'::text OR geom_ll IS NULL)
>     "enforce_srid_geom_ll" CHECK (srid(geom_ll) = 4326)
>
>
> _______________________________________________
> Qgis-user mailing list
> [hidden email]
> http://lists.osgeo.org/mailman/listinfo/qgis-user
>  

_______________________________________________
Qgis-user mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/qgis-user
Reply | Threaded
Open this post in threaded view
|

Re: QGIS performance problem

Brent Wood
Thanks Andreas,

Glad to know its not just me as well :-)

Brent


--- On Tue, 9/1/09, Andreas Neumann <[hidden email]> wrote:

> From: Andreas Neumann <[hidden email]>
> Subject: Re: [Qgis-user] QGIS performance problem
> To: [hidden email]
> Cc: [hidden email]
> Date: Tuesday, September 1, 2009, 6:31 AM
> Hi Brent,
>
> I have similar behaviour. It happens if polygons with lots
> of vertices
> are involved. Things get even worse if you try to print
> this data with
> high resolution.
>
> Marco Hugentobler and Jürgen Fischer are already working
> on this issue
> these days. They can also better explain the issues we are
> running into
> - seems to be qt and clipping related when bounding boxes
> of big
> polygons are touched when zooming in or printing on high
> resolutions.
>
> I am glad to hear that I am not the only person
> experiencing these
> problems - seems to be a real issue.
>
> Andreas
>
> [hidden email]
> wrote:
> > Hi,
> >
> > I have a PostGIS table & shapefile of the New
> Zealand coast (1:50,000)
> >
> > If I open this table it plots in a fraction of a
> second, pretty much as expected, with similar speeds for
> both the shapefile & table.
> >
> > However, if I zoom in, the redraw takes three minutes
> or more.
> >
> > While I'm waiting for the refresh, Qgis is showing
> 100% cpu load, postgres isn't showing at all (top). On an
> AMD dual core 3.2Ghz cpu with 8Gb ram it should not take
> anything like this long, and with earlier versions of Qgis
> (pre 1.0) it was much faster.
> >
> > Performance is similar with both table & shapefile
> data sources, so it does not seem Postgres related.
> >
> > However, plotting lines, points or polygons with tens
> of thousands of features rendered works well, it seems to
> primarily be polygons with many vertices (not even any
> holes) that are the problem.
> >
> > OpenSuse 11.1, Qgis 1.2.0, Postgis table described
> below, only 3 features/records, indexed & vacuumed.
> >
> >
> > Does anyone have any idea what may be the issue?
> >
> > I can provide an ssh login if anyone wants to test
> this for themselves.
> >
> >
> > Thanks,
> >
> >    Brent Wood
> >
> >
> >
> >
> >
> >
> >               
>                
>   Table "public.coast_poly"
> >  Column  |     
>    Type         
> |               
>         Modifiers     
>                
>    
> >
> ---------+-----------------------+----------------------------------------------------------
> >  gid     | integer 
>              | not
> null default nextval('coast_poly_gid_seq'::regclass)
> >  id      | bigint   
>             |
> >  fid     | bigint 
>               |
> >  name    | character varying(42) |
> >  geom_ll | geometry       
>       |
> > Indexes:
> >     "coast_poly_pkey" PRIMARY KEY,
> btree (gid)
> >     "coast_poly_geom_ll_gist" gist
> (geom_ll)
> > Check constraints:
> >     "enforce_dims_geom_ll" CHECK
> (ndims(geom_ll) = 2)
> >     "enforce_geotype_geom_ll"
> CHECK (geometrytype(geom_ll) = 'MULTIPOLYGON'::text OR
> geom_ll IS NULL)
> >     "enforce_srid_geom_ll" CHECK
> (srid(geom_ll) = 4326)
> >
> >
> > _______________________________________________
> > Qgis-user mailing list
> > [hidden email]
> > http://lists.osgeo.org/mailman/listinfo/qgis-user
> >   
>
>
_______________________________________________
Qgis-user mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/qgis-user
Reply | Threaded
Open this post in threaded view
|

AW: QGIS performance problem

Hugentobler  Marco
Hi Bernt

Yes, please provide me a remote login for debugging. I'm busy today but like to work on the freeze issue tomorrow (Wednesday). I wonder if it is really the same problem or a variation on it.

Regards,
Marco


-----Ursprüngliche Nachricht-----
Von: [hidden email] im Auftrag von [hidden email]
Gesendet: Mo 31.08.2009 20:40
An: Andreas Neumann
Cc: [hidden email]
Betreff: Re: [Qgis-user] QGIS performance problem
 
Thanks Andreas,

Glad to know its not just me as well :-)

Brent


--- On Tue, 9/1/09, Andreas Neumann <[hidden email]> wrote:

> From: Andreas Neumann <[hidden email]>
> Subject: Re: [Qgis-user] QGIS performance problem
> To: [hidden email]
> Cc: [hidden email]
> Date: Tuesday, September 1, 2009, 6:31 AM
> Hi Brent,
>
> I have similar behaviour. It happens if polygons with lots
> of vertices
> are involved. Things get even worse if you try to print
> this data with
> high resolution.
>
> Marco Hugentobler and Jürgen Fischer are already working
> on this issue
> these days. They can also better explain the issues we are
> running into
> - seems to be qt and clipping related when bounding boxes
> of big
> polygons are touched when zooming in or printing on high
> resolutions.
>
> I am glad to hear that I am not the only person
> experiencing these
> problems - seems to be a real issue.
>
> Andreas
>
> [hidden email]
> wrote:
> > Hi,
> >
> > I have a PostGIS table & shapefile of the New
> Zealand coast (1:50,000)
> >
> > If I open this table it plots in a fraction of a
> second, pretty much as expected, with similar speeds for
> both the shapefile & table.
> >
> > However, if I zoom in, the redraw takes three minutes
> or more.
> >
> > While I'm waiting for the refresh, Qgis is showing
> 100% cpu load, postgres isn't showing at all (top). On an
> AMD dual core 3.2Ghz cpu with 8Gb ram it should not take
> anything like this long, and with earlier versions of Qgis
> (pre 1.0) it was much faster.
> >
> > Performance is similar with both table & shapefile
> data sources, so it does not seem Postgres related.
> >
> > However, plotting lines, points or polygons with tens
> of thousands of features rendered works well, it seems to
> primarily be polygons with many vertices (not even any
> holes) that are the problem.
> >
> > OpenSuse 11.1, Qgis 1.2.0, Postgis table described
> below, only 3 features/records, indexed & vacuumed.
> >
> >
> > Does anyone have any idea what may be the issue?
> >
> > I can provide an ssh login if anyone wants to test
> this for themselves.
> >
> >
> > Thanks,
> >
> >    Brent Wood
> >
> >
> >
> >
> >
> >
> >               
>                
>   Table "public.coast_poly"
> >  Column  |     
>    Type         
> |               
>         Modifiers     
>                
>    
> >
> ---------+-----------------------+----------------------------------------------------------
> >  gid     | integer 
>              | not
> null default nextval('coast_poly_gid_seq'::regclass)
> >  id      | bigint   
>             |
> >  fid     | bigint 
>               |
> >  name    | character varying(42) |
> >  geom_ll | geometry       
>       |
> > Indexes:
> >     "coast_poly_pkey" PRIMARY KEY,
> btree (gid)
> >     "coast_poly_geom_ll_gist" gist
> (geom_ll)
> > Check constraints:
> >     "enforce_dims_geom_ll" CHECK
> (ndims(geom_ll) = 2)
> >     "enforce_geotype_geom_ll"
> CHECK (geometrytype(geom_ll) = 'MULTIPOLYGON'::text OR
> geom_ll IS NULL)
> >     "enforce_srid_geom_ll" CHECK
> (srid(geom_ll) = 4326)
> >
> >
> > _______________________________________________
> > Qgis-user mailing list
> > [hidden email]
> > http://lists.osgeo.org/mailman/listinfo/qgis-user
> >   
>
>
_______________________________________________
Qgis-user mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/qgis-user

_______________________________________________
Qgis-user mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/qgis-user
Reply | Threaded
Open this post in threaded view
|

Re: AW: QGIS performance problem

Brent Wood
Hmmm...

It seems that if you set up an account for Marco to try to debug this, then QGIS works perfectly...

Or it could be the OpenSuse system update I just ran, followed by a  reboot which might have fixed the problem.

QGIS is now performing admirably with lines & large/small polygons in PostGIS zooming in & out flawlessly. Both as me & as Marco.

Anyway, Marco has an account here now & I'll leave it there so either
a) it's ready for the next time, or
b) it will ensure there is no next time.


Cheers,

  Brent


--- On Tue, 9/1/09, Hugentobler Marco <[hidden email]> wrote:

> From: Hugentobler Marco <[hidden email]>
> Subject: AW: [Qgis-user] QGIS performance problem
> To: [hidden email], "Andreas Neumann" <[hidden email]>
> Cc: [hidden email]
> Date: Tuesday, September 1, 2009, 7:43 PM
> Hi Bernt
>
> Yes, please provide me a remote login for debugging. I'm
> busy today but like to work on the freeze issue tomorrow
> (Wednesday). I wonder if it is really the same problem or a
> variation on it.
>
> Regards,
> Marco
>
>
> -----Ursprüngliche Nachricht-----
> Von: [hidden email]
> im Auftrag von [hidden email]
> Gesendet: Mo 31.08.2009 20:40
> An: Andreas Neumann
> Cc: [hidden email]
> Betreff: Re: [Qgis-user] QGIS performance problem
>  
> Thanks Andreas,
>
> Glad to know its not just me as well :-)
>
> Brent
>
>
> --- On Tue, 9/1/09, Andreas Neumann <[hidden email]>
> wrote:
>
> > From: Andreas Neumann <[hidden email]>
> > Subject: Re: [Qgis-user] QGIS performance problem
> > To: [hidden email]
> > Cc: [hidden email]
> > Date: Tuesday, September 1, 2009, 6:31 AM
> > Hi Brent,
> >
> > I have similar behaviour. It happens if polygons with
> lots
> > of vertices
> > are involved. Things get even worse if you try to
> print
> > this data with
> > high resolution.
> >
> > Marco Hugentobler and Jürgen Fischer are already
> working
> > on this issue
> > these days. They can also better explain the issues we
> are
> > running into
> > - seems to be qt and clipping related when bounding
> boxes
> > of big
> > polygons are touched when zooming in or printing on
> high
> > resolutions.
> >
> > I am glad to hear that I am not the only person
> > experiencing these
> > problems - seems to be a real issue.
> >
> > Andreas
> >
> > [hidden email]
> > wrote:
> > > Hi,
> > >
> > > I have a PostGIS table & shapefile of the
> New
> > Zealand coast (1:50,000)
> > >
> > > If I open this table it plots in a fraction of a
> > second, pretty much as expected, with similar speeds
> for
> > both the shapefile & table.
> > >
> > > However, if I zoom in, the redraw takes three
> minutes
> > or more.
> > >
> > > While I'm waiting for the refresh, Qgis is
> showing
> > 100% cpu load, postgres isn't showing at all (top). On
> an
> > AMD dual core 3.2Ghz cpu with 8Gb ram it should not
> take
> > anything like this long, and with earlier versions of
> Qgis
> > (pre 1.0) it was much faster.
> > >
> > > Performance is similar with both table &
> shapefile
> > data sources, so it does not seem Postgres related.
> > >
> > > However, plotting lines, points or polygons with
> tens
> > of thousands of features rendered works well, it seems
> to
> > primarily be polygons with many vertices (not even
> any
> > holes) that are the problem.
> > >
> > > OpenSuse 11.1, Qgis 1.2.0, Postgis table
> described
> > below, only 3 features/records, indexed &
> vacuumed.
> > >
> > >
> > > Does anyone have any idea what may be the issue?
> > >
> > > I can provide an ssh login if anyone wants to
> test
> > this for themselves.
> > >
> > >
> > > Thanks,
> > >
> > >    Brent Wood
> > >
> > >
> > >
> > >
> > >
> > >
> > >               
> >                
> >   Table "public.coast_poly"
> > >  Column  |     
> >    Type         
> > |               
> >         Modifiers     
> >                
> >    
> > >
> >
> ---------+-----------------------+----------------------------------------------------------
> > >  gid     | integer 
> >              | not
> > null default nextval('coast_poly_gid_seq'::regclass)
> > >  id      | bigint   
> >             |
> > >  fid     | bigint 
> >               |
> > >  name    | character varying(42) |
> > >  geom_ll | geometry       
> >       |
> > > Indexes:
> > >     "coast_poly_pkey" PRIMARY KEY,
> > btree (gid)
> > >     "coast_poly_geom_ll_gist" gist
> > (geom_ll)
> > > Check constraints:
> > >     "enforce_dims_geom_ll" CHECK
> > (ndims(geom_ll) = 2)
> > >     "enforce_geotype_geom_ll"
> > CHECK (geometrytype(geom_ll) = 'MULTIPOLYGON'::text
> OR
> > geom_ll IS NULL)
> > >     "enforce_srid_geom_ll" CHECK
> > (srid(geom_ll) = 4326)
> > >
> > >
> > > _______________________________________________
> > > Qgis-user mailing list
> > > [hidden email]
> > > http://lists.osgeo.org/mailman/listinfo/qgis-user
> > >   
> >
> >
> _______________________________________________
> Qgis-user mailing list
> [hidden email]
> http://lists.osgeo.org/mailman/listinfo/qgis-user
>
>
_______________________________________________
Qgis-user mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/qgis-user