Hello,

The release notes for LLBLGen Pro v2.0 say that one of the improvements is a greatly enhanced fetch speed. Can you give figures on this improved speed? We are investigating the upgrade to a new version of llbl because of performance improvements. But the changes in the template system prevents us from doing some quick tests. So we first want to have an indication on the improvement before spending time on migrating our templates. I know that performance is related to the type of the queries, but maybe you can give some figures for specific kinds of queries.

Regards, Marco Zuiderwijk

Memory consumption is way lower, think 30-40% lower. Fetch speed of queries without dependency injection in v2.5 is much much faster than 1.0.2005.1, I think 40-50% faster or thereabout. You also have much more tweakability: you can exclude fields now if you don't need them, which can enhance performance in hotspots if you fetch entities but don't need big blob fields at that point for example.

v1.0.2005.1 was slower because of the inheritance code implementation. Inheritance implementations slow down the queries a lot, relatively speaking. For example 1.0.2004.2 is much faster. In v2.0 we eliminated a lot of bottlenecks in the code, and in v2.5 we optimized even more elements.

There's no way we can optimize it even further without generating IL at runtime.