Hi all,
I just wanted to share with everyone that my app engine site is now
live with brand new pyTenjin templates. Previously I had been using
the built-in django templates but converted in an effort to reduce CPU
usage, particularly on uncached templates.
In case you're not familiar with google app engine, it automatically
spawns runtimes to service concurrent requests and then tears down
instances after several seconds of inactivity. While the built-in
django templates are cached in memory for a given runtime, they are
read and compiled fresh for each new runtime spawn.
On my most complicated template uncached requests' processing time was
reduced 10-15% using pyTenjin. With pyTenjin templates can also be
cached via the memcache service. Requests which utilize in-instance
memory cached templates took approximately 30% less time with
pyTenjin. Note that those percentages are for full request processing
which involves a lot of data retrieval. The percentage improvement in
template compiling and rendering alone is necessarily higher.
I'd like to especially thank Makoto Kuwata for all the help with
making helper functions and template subclassing. Without all the
help, porting my django templates would have been much more difficult.
If you'd like to see the site, it's live at
http://www.myweightracker.com
but, if everything is working as it should, you'll never see Tenjin as
it works quietly behind the scene.
Cheers,
Steve