Bug 28225 - very slow rendering
Summary: very slow rendering
Status: RESOLVED MOVED
Alias: None
Product: poppler
Classification: Unclassified
Component: general (show other bugs)
Version: unspecified
Hardware: Other All
: medium normal
Assignee: poppler-bugs
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-05-24 02:00 UTC by Johannes Berg
Modified: 2018-08-20 22:08 UTC (History)
2 users (show)

See Also:
i915 platform:
i915 features:


Attachments
file that is very slow to render (820.58 KB, application/pdf)
2010-05-24 02:00 UTC, Johannes Berg
Details
Rendering of this pdf is extremely slow (114.99 KB, application/pdf)
2012-03-23 14:55 UTC, thomas
Details
Reduced image resolutions to 600 ppi with acrobat x (184.75 KB, image/tiff)
2012-05-14 06:21 UTC, Thomas Freitag
Details
Generated PDF using TikZ that renders REALLY slowly (16.05 KB, application/octet-stream)
2013-11-02 21:27 UTC, Adrià Arrufat
Details

Description Johannes Berg 2010-05-24 02:00:09 UTC
Created attachment 35813 [details]
file that is very slow to render

The file at http://www.strassentheater-detmold.de/fileadmin/PDF_DATEN/Programm_nach_Tagen_FH_ENG_nach_%C3%84nderung.pdf takes a very long time to render. Maybe this is a cairo problem?

This file also takes forever to render on my N900 which I think is based on xpdf directly, so might be an old issue?

This is poppler version 0.8.7-3 (debian) I think.

I have also attached the file just in case it disappears from the website.
Comment 1 Albert Astals Cid 2010-05-25 15:55:33 UTC
Not evince specific
Comment 2 thomas 2012-03-23 14:55:17 UTC
Created attachment 58950 [details]
Rendering of this pdf is extremely slow
Comment 3 thomas 2012-03-23 14:59:44 UTC
I'm sorry if I should have opened a separate bug for this; I don't know how to
check whether it's related or not.

The pdf I have here (see attachment) renders _extremely_ slowly (evince &
okular; the file attached by the OP renders blazingly fast compared to this
one).

poppler version (openSUSE 12.1) is 0.18.0
Comment 4 Albert Astals Cid 2012-03-24 08:20:07 UTC
Thanks for the new pdf. But next time, if you don't know if it's the same bug or not, please don't hijack other people bugs, it's much harder to properly mark it as fixed when there is random pdf attached to it that have nothing to do with eachother.
Comment 5 Albert Astals Cid 2012-03-24 08:23:01 UTC
And yes, this pdf has nothing to do, so please open a separate bug.
Comment 6 thomas 2012-03-29 08:57:14 UTC
Sorry again, I'll remember it for future reports. Separate bug report is here:

https://bugs.freedesktop.org/show_bug.cgi?id=48051
Comment 7 Thomas Freitag 2012-05-02 00:02:00 UTC
In the actual version under ubuntu, 64 bit, it renders in 13 seconds, so 4-5 seconds per page. That seems to be a resonable time for me. And I can't see any problem with the PDF of comment 2 (s. bug 48051, too)
Comment 8 Albert Astals Cid 2012-05-10 10:12:02 UTC
I get 3seconds per page in my laptop, but to be honest seems a bit too much for a "simple" page as that (text and some small images) i'd be interested in knowing why it's so slow
Comment 9 Thomas Freitag 2012-05-14 06:16:11 UTC
(In reply to comment #8)
> I get 3seconds per page in my laptop, but to be honest seems a bit too much for
> a "simple" page as that (text and some small images) i'd be interested in
> knowing why it's so slow

It's not just so "simple" You think: all the yellow symbols are images with a resolution of 4924.5 ppi, and because we don't cache xobjects, we read i.e. on page 1 70 images with such a resolution. This takes some time, even if the images are all jpeg compressed, so one of the compressions which can be really read quickly.
Of course it would be quite easy to implement a caching for images. If we would do it, we would read i.e. the Euro sign on the first page only once instead of 8 times. But the next PDF producer probably even doesn't recognize that is is always the same sign and would insert 8 images instead of one, and then even a cache will not help anyway.
Therefore I think that 3 seconds per page is acceptable for something like this.
Comment 10 Thomas Freitag 2012-05-14 06:21:29 UTC
Created attachment 61613 [details]
Reduced image resolutions to 600 ppi with acrobat x

I optimized the PDF with acrobat x. It has still a quite huge quality, i.e. can be zoomed highly before You begin to see pixels. But if You think that this is a common problem I would implement a cache some day...
Comment 11 Adrià Arrufat 2013-11-02 21:27:17 UTC
Created attachment 88538 [details]
Generated PDF using TikZ that renders REALLY slowly

The other day I generated a PDF using LaTeX TikZ that renders really slowly in Evince (not in Okular or Mupdf...)
So I guess it's a poppler-glib specific bug...

In the attachment I include the PDF and the LaTeX sources
Comment 12 Programmist11180 2014-01-25 13:20:02 UTC
I have the same issue. File from first attachment renders very slow.

Also, I found file with same problem. You can find it here http://dfiles.ru/files/2r1qpx7xm

Sometimes render of page goes into infinite loop and never ends.

I use qpdfview 0.4.7, poppler 0.24.5-1 on Debian.
Comment 13 Albert Astals Cid 2014-09-21 10:23:29 UTC
Adrià, don't play with bug importance.
Comment 14 Adrià Arrufat 2014-09-21 10:27:00 UTC
(In reply to comment #13)
> Adrià, don't play with bug importance.

Sorry, it was a test, I didn't know anyone could modify the bug importance :(
Comment 15 Adrià Arrufat 2014-09-21 10:35:21 UTC
By the way I was wrong about the okular not bugging... It renders the file I attached much faster than evince, but once you zoom, there's a huge memory leak
Comment 16 Albert Astals Cid 2014-09-21 10:43:52 UTC
You'll have to give proof it's a memory leak with a valgrind run output, otherwise i'm not keen of accepting people saying "it's a memory leak!"
Comment 17 GitLab Migration User 2018-08-20 22:08:24 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/poppler/poppler/issues/209.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.