jastrachan | 15 May 08:09 2004

Re: Groovy Performance 20-90% of Java?

On 14 May 2004, at 21:13, Walter Barrett wrote:
> I just reviewed Rod Cope's presentation at
> http://www.openlogic.com/presentations/
> According to the following slide,
> http://www.openlogic.com/presentations/groovy/groovy_pres-djug 
> -05122004_files/frame.htm
> Groovy's "performance is 20-90% of Java depending on usage."
> 20-90% is significant in my opinion. Do others feel this weak  
> performance is accurate?

Note that we've not really spent much time tuning the implementation of  
Groovy to be high performance. For example when static typing is used,  
there's no reason why the groovy compiler can't output pretty much  
exactly the same bytecode as javac; so there would be no performance  
loss whatsoever. For ease of implementation, we're using dynamic method  
dispatch for all methods.

Dynamic method dispatching does have a performance cost, though I'm  
sure over time we can tune the implementation to make it a fairly minor