Speed of Java (was: Linux on old laptop in two stages)

Christopher Chisholm christopher.chisholm at syamsoftware.com
Tue Jun 6 14:29:00 EDT 2006


I've always wondered:

could you build a distro of linux that has native java support?  when 
you think about how java works, it compiles programs to byte code, which 
is then executed on each platform by a virtual machine.  what if someone 
created a linux kernel that could directly interpret and execute this 
byte code?  programs could have direct access to system resources, and 
it wouldn't be up to a virtual machine to handle memory management and 
other areas that affect efficiency. 

my biggest problem with java is that every java app i use FEELS bulky, 
even on a crazy high-end desktop.  practically, they run just fine.  but 
i don't want to wait an extra half-second damnit.  nothing feels as nice 
as a UI written in C/C++ and powered by either openGL or native 
windowing methods of an OS.  Java is great to work in since the java 
library has methods for practically everything (DishWasher dishwasher = 
new DishWasher.WashMyDishes()), but I can never get used to the laggy feel.

-chris

Ben Scott wrote:
> On 6/6/06, Lawrence Tilly <mail.list.tilly at gmail.com> wrote:
>> http://wiki.java.net/bin/view/Games/JeffOnPerformance
>
>  There seem to be an awful lot of exceptions and conditions in there.
>
>  For example, the author states you should have a "modern" JIT
> compiler (whatever "modern" means), and ignore all the time at the
> start where JIT optimization hasn't happened yet and is in fact making
> things slower.  Well, that's sure convenient.  Many of the Java
> programs I have to use, I only use for brief periods of time -- that
> same time I'm supposed to ignore.  So I guess I should ignore Java
> entirely!  ;-)
>
>  Likewise, the "J2ME/CLDC" exception.  The author never specifies
> what "slow" and "very limited" mean.  From what I gather from Java
> people here on *this* list, it sounds like less than 512 megabytes is
> "limited" in Java terms.  ;-)
>
>  I also noticed that the benchmarks the author links to are focused
> on "math heavy" operations.  All the Java programs I have to use
> aren't doing lots of math, but rather, things like string
> manipulation, network protocol, systems management, databases, etc.
> Maybe that's significant.
>
>  The other Java stuff I have to use is the random Java applets I
> encounter on the web.  And, without fail, every time I encounter one,
> the browser grinds to a halt.  Blame it on whatever you want, but
> that's real world experience where Java == slow.  Doze and Nix both.
> Firefox and MSIE.  MS-VM and Sun JRE.  Maybe there's a fast JVM out
> there that all the smart Java people are using, but for everyone else,
> it sucks mud.
>
>  I really find this a shame, as Java-the-language seemed pretty nice
> to me when I was exposed to it years ago.
>
> -- Ben
> _______________________________________________
> gnhlug-discuss mailing list
> gnhlug-discuss at mail.gnhlug.org
> http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss
>
>
>




More information about the gnhlug-discuss mailing list