Last modified: 2013-07-04 10:34:23 UTC

Wikimedia Bugzilla is closed!

Wikimedia migrated from Bugzilla to Phabricator. Bug reports are handled in Wikimedia Phabricator.
This static website is read-only and for historical purposes. It is not possible to log in and except for displaying bug reports and their history, links might be broken. See T42577, the corresponding Phabricator task for complete and up-to-date bug report information.
Bug 40577 - Parsoid: Possible performance regression on latest master
Parsoid: Possible performance regression on latest master
Status: RESOLVED FIXED
Product: Parsoid
Classification: Unclassified
General (Other open bugs)
unspecified
All All
: Low normal
: ---
Assigned To: Gabriel Wicke
: performance
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2012-09-27 18:18 UTC by Gabriel Wicke
Modified: 2013-07-04 10:34 UTC (History)
1 user (show)

See Also:
Web browser: ---
Mobile Platform: ---
Assignee Huggle Beta Tester: ---


Attachments

Description Gabriel Wicke 2012-09-27 18:18:44 UTC
[[:en:San Francisco]] takes 151 seconds to render, which is way too long.

We should check if this is a regression (by trying with an earlier version) and do some profiling to identify the expensive bits in this page.
Comment 1 Gabriel Wicke 2012-09-27 18:26:53 UTC
[[:en:Barack Obama]] went from 58 seconds a few weeks ago to 76 seconds on my laptop, so performance seems to have deteriorated quite a bit.
Comment 2 Gabriel Wicke 2012-09-27 21:53:07 UTC
The SF page seems to be problematic in the PHP parser too:

[14:14] <brion> bah Fatal error: Maximum function nesting level of '100' reached, aborting! in /srv/com.leuksman/mw/includes/parser/Preprocessor_DOM.php on line 985
[14:14] <brion> rendering copy of 'San Francisco'
[14:14] <brion> sighhhhh
Comment 3 ssastry 2012-09-28 00:31:42 UTC
Here is a result of timing runs for the Barack Obama page on my laptop (node version 0.6.12).  I did 5 runs and picked the best value to minimize network latency effects on my home wifi.  Overall, nothing alarming that I see here.

+---------+------+------+------+------+------+------+---------------------------------------+
| Revision| BEST | r1   | r2   | r3   | r4   | r5   | Revision detail                       |
+---------+------+------+------+------+------+------+---------------------------------------+
| master  | 51.8 | 57.3 | 51.8 | 54.3 | 53.6 | 52.0 | category rt fix (and related patches) |
| 78cf1ce | 52.1 | 54.5 | 52.1 | 52.7 | 53.7 | 53.1 | ref tokenizer fix                     |
| d12687d | 52.7 | 54.1 | 56.7 | 53.9 | 52.7 | 54.2 | pre-ref tokenizer fix                 |
| 53e823a | 51.0 | 51.0 | 51.7 | 51.4 | 52.9 | 53.9 | templ-encap bug fix                   |
| 9ea500c | 51.6 | 52.0 | 52.7 | 52.2 | 51.6 | 53.5 | fix regression in tpl-affected attrs  |
| 9b91f43 | 50.4 | 52.3 | 50.4 | 51.1 | 52.3 | 54.0 | dont throw away useful dsr info       |
| 7ed044e | 49.5 | 49.5 | 50.0 | 50.1 | 51.6 | 50.6 | p-wrapping                            |
| af8a6fa | 48.2 | 48.7 | 49.0 | 48.2 | 50.4 | 49.5 | case-insensitive link handling        |
+---------+------+------+------+------+------+------+---------------------------------------+
Comment 4 Gabriel Wicke 2012-09-28 04:37:01 UTC
Yeah, just a steady slowdown in these revisions, and not unexpected for the extra work on dsr calculations and template round-tripping.

Decreasing the priority for now, can look for optimization opportunities with node profiler later.
Comment 5 Gabriel Wicke 2013-05-15 18:52:27 UTC
This is now really old, no need to look into this any more.
Comment 6 Andre Klapper 2013-07-04 10:34:23 UTC
[Parsoid component reorg by merging JS/General and General. See bug 50685 for more information. Filter bugmail on this comment. parsoidreorg20130704]

Note You need to log in before you can comment on or make changes to this bug.


Navigation
Links