[rc5] Re: rc5-digest V1 #96

Michael Chajkowski mickyc at superiway.net
Mon Aug 25 15:26:26 EDT 1997


Ok, now I do NOT understand what all this griping about the speed of
Java is about!!!  A reasonable structure for the client has already
been suggested.  It goes like this:

A Java application ( NOT APPLET ) that does all disk and network I/O,
handeling the administration of new and finished blocks, as well as
providing a common GUI.  Now, the real work gets done!  Initially
program the actual code cracking mechanism in C, C++, or whatever,
write it for maximum compatability, so that the port to any CPU can be
done fairly quickly (read: just recompile in many instances).  This can
all be done fairly quickly.  Once this is finished and the client is
stable, rewrite the cracking code, and ONLY the cracking code, in
assembler.

This would provide a common look and feel to the program, while
generating the optimum performance for the actual cracking code.  It
would also speed up initial development time, partially becauase you
have to code the user interface only once, and initially compile the
code cracker only once for each CPU!!!

On Mon, 25 Aug 1997 11:15:33 -0400, rc5-digest wrote:

>>> Sigh, Java code is slower than taking the v1 source and compling it on
>>> your own machine.  In fact, it is 130 times slower, and due to "features"
>>> in Java, you would have to have a proxy server on the server you retrieved
>>> the url from.  It is a nice idea, and it was done, and it was slower than
>>> a SE/30.  Since the machines that Java has been ported to already have
>>> native client, the loss in performance would be tremendous.
>>
>>Agreed. I think most people would be put off by the speed penalty of Java.
>
>What a pissy trade-off.  The huge step that one source would provide vs. 
>the slowness of java.
>
>However, the JIT compilers have been coming into fruition.  The web site
>
>http://www.intergalact.com/java/speedtest/speedtest.html
>
>suggests that JIT compilers can be faster than C++.  Of course this will 
>vary greatly, but it's encouraging.  Compiled Java will never be as fast 
>as assembly, but is in the neighborhood.
>
>Anyway, a Java based client is not impossible and the speed lost by not 
>going to assembly can be made up by more people running the client.
>
>Another idea:  There is a lag in writing the different platform clients, 
>Java could be the bridge to start a new project on all platforms and then 
>swap in the native code when it gets done.

******************************
Michael Chajkowski
Desktop Technician
StarField Productions
email: mickyc at superiway.net
------------------------------
A Warped mind is a terrible thing to waste...
   Use OS/2 Warp!

If you are not using OS/2 Warp, you have my sympathy.
******************************

If you would like my PGP key, send me an email
with "Get-PGP-Key" as the subject.

----
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.



More information about the rc5 mailing list