performance improvements #519
Conversation
…site creation up of a 20%.
…ayload in post render to reduce the number of calls
|
👍 What version of ruby were you benchmarking with? |
|
I am using 1.9.2p290(on a MacBookAir 2.13Ghz Core 2 Duo 4GB RAM). |
| @@ -183,10 +183,11 @@ def related_posts(posts) | |||
| # Returns nothing | |||
| def render(layouts, site_payload) | |||
| # construct payload | |||
| sp = site_payload | |||
There was a problem hiding this comment.
Isn't site_payload already a local variable here? It's passed into the function.
There was a problem hiding this comment.
yes, you re totally right.
my bad.
I'm going to revert this one fix in my fork.
Dp you want a new pull request ?
There was a problem hiding this comment.
You can force push to this pull request's branch.
There was a problem hiding this comment.
Thanks.
I can actually see my last commit on my fork reverting that code down here in this pull request thread.
I guess it's now part of the pull request then.
… into the method. late night mistake :-P
|
hey I can see that the pull request has been merged in master (right now it's the last commit). Is then the right time to close it as well ? |
|
Merged in 7d88f72. |
I did some profiling and noticed how jekyll spends a lot of time in doing comparisons, namely Post comparison (via <=>).
Post comparison is actually mostly called from site_payload, that generates a on-demand hash containing (among others) the ordered list of posts. In methods where site_payload is called more than once I just introduced a local variable to reduce the multiple calls and the additional comparison.
Just this in my benchmark gives an improvement in the time needed to generate the site of a 20-25%(particularly useful for big site, with over 1000 between pages and posts).