[Jython-users] slow JPython math on HP??

[John M. <joh...@ya...>]

Why is math sooooooooo slow on HP versus Linux??  For the example below
I calculate the time to calculate 2L**10000 on HP and Linux.  It's over
2,000 faster on a Linux PC (512 MHz dual Celeron) than on a HP (K box,
4 processor, 180 MHz).  

This particular example is not critical to my application.  I just
focus on it because of the dramatic difference in performance on HP
versus other platforms.  Might this be a clue as to a problem with HP??


HP:

$ jpython
JPython 1.1+09 on java1.2.2.05 (JIT: null)
Copyright (C) 1997-1999 Corporation for National Research Initiatives
>>> from time import time
>>> t1=time(); x=2L**10000; delay=time()-t1; print delay
86.5       <--  interpretted, very slow
>>>
$ jpython
JPython 1.1+09 on java1.2.2.05 (JIT: HP)
Copyright (C) 1997-1999 Corporation for National Research Initiatives
>>> from time import time
>>> t1=time(); x=2L**10000; delay=time()-t1; print delay
11.539999961853027
>>> t1=time(); x=2L**10000; delay=time()-t1; print delay
11.8     <-- classic JIT
>>>
$ jpython
JPython 1.1+09 on java1.2.2.05 (JIT: null)
Copyright (C) 1997-1999 Corporation for National Research Initiatives
>>> from time import time
>>> t1=time(); x=2L**10000; delay=time()-t1; print delay
14.9     <--  HotSpot JIT
>>> $ ^C
$




Linux        512 MHz   dual Celeron (brain dead?) Pentium

 Linux: /home/mudd[1]  jpython
JPython 1.1 on java1.2.2-RC2 (JIT: javacomp)
Copyright (C) 1997-1999 Corporation for National Research Initiatives
>>> from time import time
>>> t1=time(); x=2L**10000; delay=time()-t1; print delay
0.00499        <-- About 2,400 times faster than HP's best numbers??!?
>>>
 Linux: /home/mudd[1]




__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail. 
http://personal.mail.yahoo.com/