Well, I just ran it on a 233 MHz Pentium with the appletviewer, JDK 1.1.7v1a and
TYA 1.2, and got a score of 13.0, while according to the table, the performance
of the Dell Optiplex 200 MHz P6, Win 95, Netscape 4.04 is 42.5 (about three
times). I don't do Windows, so I can't tell you what the results are on my
machine with WinBugs...
Now, while the results are certainly not earth-shattering, I've used about 4
versions of TYA and seen (subjectively) increasingly better performance, so you
can expect that to improve. Also, I develop with Java under Linux and it's
*DEFINITELY* acceptable, especially if you use Jikes to compile. The runtime
performance is also good (subjectively), so I'd say, "Go for it!"
-Mario.
"Marvin McNett II (GTA)" wrote:
> I recently ran the Java benchmark at:
>
> http://math.nist.gov/scimark/
>
> on my 300MHz PII Linux machine with Netscape 4.5. To my astonishment, the
> results were terrible! Since I can dual boot, I started up Windows 95 (a
> rare occurance) and ran the benchmark again (and again with Netscape 4.5).
> I found that the score was about 20 times higher (higher is better) using
> Windows 95! So, I concluded that Netscape for Linux must not have a
> just-in-time (JIT) compiler, whereas Netscape for Windows does. To verify
> this, I, once again, ran the test in Linux--this time using JDK 1.1.6. I
> I found that the Windows 95 score was still about 8 times higher. So, I'm
> wondering, "Is there a problem with Java on Linux?". Am I doing something
> wrong? I need to use Java for an upcoming project, but certainly do not
> want to use Micro$oft.
>
> Thank you,
>
> Marvin McNett II
