Should really execution performance be the "selling point" ? I would really 
rather try to convince on better usability/maintainability for the 
developer, faster development cycles, all that with stable performances 
(rather than "best" performances).
Otherwise, in case you have missed them, some usually "easy" performance 
boost in Nhibernate : enabling batching on lazy loaded collections / entity 
properties (to avoid n+1 performance issues), avoiding fetch joins (to 
avoid loading unneeded data), maybe configuring adonet.batch_size. Of 
course this should not be considered as decisions to always take, it still 
depends on your use cases.

Le vendredi 21 décembre 2012 00:14:23 UTC+1, Randar Puust a écrit :
>
> So I'm a huge fan of NHibernate and trying to move our current platform 
> from Entity Framework 4.0 to NHibernate.  Unfortunately, I just ran into a 
> problem, which is that the performance seems much worse when we tested it 
> in an isolated set of tests, using the same patterns we currently use.  Of 
> course my concern is that to win the NH argument, I have to show solid 
> evidence that NH will give us a performance boost, along with all the other 
> things I love so much about NH.  
>
> My baseline is EF 4.0 with Model First.  The tests included operations 
> like Create, Publish and Edit, which are specific to our application.
>
>    - EF 5.0 and Code First – 166% slower
>    - EF 5.0 and Model First with .Net 4.5 – 100.8% slower
>    - EF 5.0 and Code First with .Net 4.5 – 171% slower
>    - NHibernate 3.3.1 – 163% slower
>
> So based on this, my best solution is just to stick with Model First and 
> probably upgrade to 4.5.  What I'm really looking for is some good 
> performance data that compares EF 5.0 to NH.  The best I can find is this, 
> but it's from 2009
>
>
> http://gregdoesit.com/2009/08/nhibernate-vs-entity-framework-a-performance-test/
> I really don't think I'm going to be able to convince people to re-factor 
> the entire DAL without more solid evidence that NH is on average faster.  
> Does anybody know of any analysis that has been done?
>

-- 
You received this message because you are subscribed to the Google Groups 
"nhusers" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/nhusers?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to