[plt-scheme] Quick comparison MzScheme3m vs Scsh

From: Jaime Vargas (jev at mac.com)
Date: Fri Aug 4 16:50:58 EDT 2006

I was deciding on which scheme implementation to use for an embedded  
project.
I ran a quick test to see the cpu and ram footprints of MzScheme3m  
and Scsh,
here are the results:

Resources used at launch time.

   PID USERNAME PRI NICE  SIZE   RES STATE    WAIT     TIME    CPU  
COMMAND
  6571 jvargas    2    0 4384K 3644K sleep    select   0:00  0.00%  
mzscheme3m
10652 jvargas    2    0   20M 5188K sleep    select   0:00  0.00% scshvm

Now lets do some testing, how about the factorial of 10000. I am  
going to
use a (apply * (iota 10000 1)). In mzscheme I need to load srfi-1, which
increases the footprint a bit.

   PID USERNAME PRI NICE  SIZE   RES STATE    WAIT     TIME    CPU  
COMMAND
  6571 jvargas    2    0 5908K 4348K idle     select   0:00  0.00%  
mzscheme3m

Lets see what happens when we run the test.

   PID USERNAME PRI NICE  SIZE   RES STATE    WAIT     TIME    CPU  
COMMAND
  6571 jvargas    2    0 8032K 7308K sleep    select   0:01  2.49%  
mzscheme3m
29908 jvargas   64    0   19M   20M run      -       18:22 99.02% scshvm

Not only mzscheme3m use less ram, it also finish a lot faster, 1100  
times
faster. This is a pleasant surprise! And I now know which tool to use  
for
my project.

I know this is not a really serious benchmark, but just the small memory
requirements of mzscheme are enough for my decision. Kudos to the PLT  
team!

--
Jaime


Posted on the users mailing list.