On Mar 4, 9:08 am, gvkalra <gvka...@gmail.com> wrote: > hi .... i have a presentation to make on django, wherein I have > decided to include a section on Tornado .... I wish to show > performance comparisons between django and tornado ..... how do I set > up the scenario ?? ..... what I intend to achieve is this > graph:http://wiki.developers.facebook.com/images/e/e8/Chart.png.... Please > help me out
You do understand that those original Tornado benchmarks are to a degree bogus as far as they were being used to compare Tornado to Apache/mod_wsgi? If you do a legitimate comparison and compare their basic Tornado hello world program to a basic WSGI hello world program then Tornado doesn't come out looking so good. On MacOS X, Tornado actually performs worse than Apache/mod_wsgi for a basic hello world in each system. In other words, they should not have been comparing the performance based on running a heavy weight Python framework like Django on top of Apache/mod_wsgi to a native hello world program in Tornado. A more realistic and fair comparison would be to compare Django running on top of Tornado to Django running on top of Apache/mod_wsgi. One of the things you will find from that is that any benefits from Tornado being asynchronous get thrown out the window as to run a WSGI application such as Django on top of Tornado you need to implement a pool of blocking threads on top and as soon as you do that, the overall performance of Tornado will drop. That this occurs is because once you get a blocking WSGI model in play, the bottleneck is usually never going to be the underlying hosting mechanism and instead the bottleneck is your application and database access. So, what are you trying to show? Are you wanting to show that a blocking Django or WSGI application running on Tornado is faster? If you are, then you will have trouble doing that. If instead you want to show that the asynchronous model of web programming gives you an advantage for your specific application, then why use Django in the comparison at all. In that case it would be more realistic to compare custom low level WSGI applications to a custom low level Tornado specific application. Help us to understand what result you want, whether that be positive to Tornado or Apache/mod_wsgi, or somehow show the benefits of Django for web programming, then there is no doubt a way of slanting the results to get what you want. Graham -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this group, send email to django-us...@googlegroups.com. To unsubscribe from this group, send email to django-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/django-users?hl=en.
