Django has low benchmark
Gary
Etienne Robillard
Disclaimer: I'm the author of the Django-hotsauce toolkit, a high-performance and scalable fork of Django.
This result doesn't surprise me. I would love to see someone make benchmarks between Django-hotsauce and pure Django.
I believe Django is trying to solve too many problems at once, and may not be optimized for high-performance WSGI applications.
In specific, perhaps Django is way too monolithic and not modular enough to challenge specialized web microframeworks like Flask or Werkzeug.
Hence, I would love to have the ability to disable specific modules in Django which I don't need.
For instance, I don't use the ORM or template code in my wsgi applications.
I switched to Mako templates and rolled my own model API to reduce this overhead.
HTH
Etienne
--
You received this message because you are subscribed to the Google Groups "Django users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to django-users...@googlegroups.com.
To post to this group, send email to django...@googlegroups.com.
Visit this group at https://groups.google.com/group/django-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/django-users/11a21a82-4dbe-4515-babe-1e9d468e0f35%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
-- Etienne Robillard tka...@yandex.com https://www.isotopesoftware.ca/
Antonis Christofides
Hello Gary,
To me these comparisons are largely meaningless. For example, they compare PostgreSQL to MongoDB. You can't really do that any more than you can compare an apple to an orange. They compare aiohttp to Django; my first reaction to that was "why is aiohttp only twice as fast as Django?". On how much RAM is this? What is the PostgreSQL client library they use for aiohttp? Is it async? What's the point of using aiohttp for a database intensive application if you don't have an async PostgreSQL client library? Is it aiohttp with the standard library loop or with uvloop? Likewise with Django, they use some kind of asynchronous wsgi server, but what about the PostgreSQL client library? Assuming it's synchronous, how many processes do they run?
Even if these questions are answered, we'd have more questions. What application is this for? Why was it made this way and not in another way? Why does it not use a cache or why does it use a cache?
Why do you care about these benchmarks? What is the problem you are trying to solve by consulting these benchmarks?
Regards,
Antonis
Antonis Christofides http://djangodeployment.com