Hi, i'm author of the snippet http://www.djangosnippets.org/snippets/507/. I have used simple dictionary because pickling and unpickling values from the cache are IMHO to expensive and don't give a big speed boost over parsing. Also if you use memcache or filesystem backend for caching (pretty standard thing in production environment) you don't really have a speed boost because you have to use IO to get template (+pickle).
In my tests with pretty simple templates, lighttpd and fastcgi django process, with my setup i've got from ~400 rps to almost 900 rps. You could just create a blank app in your project and paste my snippet in __init__.py file of that app, and put that app into your INSTALLED_APPS tuple. Now you can enable and disable template caching with just one line in your settings. On Dec 23, 8:00 am, "Rob Hudson" <[EMAIL PROTECTED]> wrote: > On 12/21/07, Malcolm Tredinnick <[EMAIL PROTECTED]> wrote: > > > In the early days of Django, Adrian, Simon et al looked at that. It > > wasn't worth it, since, in the grand scheme of things, template caching > > and checking the cache wasn't that much faster than loading and parsing, > > particularly in the overall response time of a request (of which > > template parsing is a relatively small component). Adding complexity for > > minimal gain isn't usually a good idea. Unless this is a universal win, > > it would be better to write it as a third-party template loader. It's > > fairly easy to write a template loader that takes another template > > loader as a parameter and just wraps caching around it and that keeps > > the core code cleaner. > > I wonder if the template system has become a bit more complex since > then. I also wonder if whatever tests they used included things like > includes in for loops. I tend to think that the filesystem is slow > and anything to remove FS calls and shove things in memory is a good > thing -- especially something that could potentially be in a for loop. > Obviously there are trade-offs as you mention but the patch didn't > look that complex to me -- actually it was surprisingly straight > forward. > > I've considered applying this patch and testing against a project I'm > working on. Maybe that would help prove to either Django or me that > this is or isn't worth it. > > Thanks, > -Rob --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Django users" group. To post to this group, send email to django-users@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-users?hl=en -~----------~----~----~----~------~----~------~--~---
