[Mono-list] Trolltech Qt 4 for MONO

[Jiří Zárevúcký]

2009/2/25 David Canar <davidcanar at gmail.com>:
> On Tue, Feb 24, 2009 at 7:57 PM, Jiří Zárevúcký
> <zarevucky.jiri at gmail.com> wrote:
>>> qt4dotnet is a slightly modified Qt Jambi compiled using ikvm
>>
>> Problem 1: It's a Java code running in Java VM which is running on Mono.
>>
>> I don't know, how does IKVM performs in the terms of performance or
>> memory usage, so correct me if I'm wrong, but isn't that utterly
>> ineffective?
>
> Not at all. I haven't made any formal performance comparisons but you
> can't notice any difference between a Qt C# app and a QtJambi app
> (time to load, speed, responsiveness, and even memory consumption,
> this is talking about Microsoft's .NET; mono still has some unresolved
> issues throwing exceptions that make it slower). But since you
> mentioned it I should make some formal tests comparing between them
> and see how "ineffective" this approach is. Regardless of it the
> simple fact that you can get .NET bindings for the whole Qt 4 stack in
> 5 minutes I think is a huge plus in "effectiveness".
>

Thanks for response. :)

Perhaps I was just misguided on how IKVM works.

http://www.mono-project.com/Java says:
"There are two possible ways of using IKVM: one is to use it as a
Just-in-Time compiler which translates the Java bytecodes into .NET
Intermediate Language as it goes. But this means that at runtime you
are compiling things twice: the Java-to-CIL JIT and the CIL-to-Native
JIT."

The second way described is AOT compiling of the .NET wrapper to
native code, not the Java itself to IL.

Is there something incorrect, or is this overhead simply not noticeable?