'Fast' dependency injection? about dagger

[patrickbaumann]

It's nothing official or scientific, but I threw together a couple of demo Android apps that are identical in function but built with RoboGuice in one implementation and Dagger in the other. Source and results are at http://github.com/patrickbaumann/daggervsroboguice

It's a simple example but in the worst case, Dagger sliced off about 350ms of start up time compared to Guice on a Galaxy Nexus.

[scompt]

Here's a tweet from @swankjesse where he mentions a couple numbers between Guice and Dagger.

[swankjesse]

On Fri, Dec 7, 2012 at 1:14 AM, Joe [email protected] wrote:

Are there any plans to benchmark or put numbers behind the 'fast' claim?
I'm curious how it compares to Guice specifically.

Yeah, the speed difference will depend on the size of your app. Suppose you
have a small app that spends 100ms in Guice each time it starts up. The
maximum improvement you could see in such an app is 100ms. Some of the apps
I've worked on spent a long time in Guice at startup: 3000ms or more (on
some older devices, anyway). This is where Dagger's avoidance of reflection
and runtime annotations pays off.

[JakeWharton]

Closing. For further discussions please see the Google+ community or the mailing list.