Benchmarks updated

Earlier I posted some rather non technical benchmarking result. Also, the python times was compared in the interactive console. Here are the results running from the command line:

IronLisp: Build in release mode
Z:\dev\IronLisp\IronLisp\bin\Release>IronLisp.exe
time: 326.281ms
time: 303.722ms
time: 313.096ms
time: 304.353ms
time: 305.102ms

IronPython: Latest checkout, build in release mode
C:\dev\IronPython_Main\Bin\Release>ipy tak.py
Result: 7 Time: 0.763377747731
Result: 7 Time: 0.657344718393
Result: 7 Time: 0.655193607009
Result: 7 Time: 0.664234979585

Python 2.51
C:\Python25>python.exe -O tak.py
Result: 7 Time: 1.38913744624
Result: 7 Time: 1.38743890634
Result: 7 Time: 1.37719849869
Result: 7 Time: 1.37393355859

As you can see this paints a different picture. The picture is indeed very good as the DLR compiles most of the code. I tried running IronPython in ‘eval’ mode, but it seem to think about something forever 🙂 (this didnt happen though previously when tested on the interactive console).

Here is the python code:

import time

def tak (x, y, z):
    if not(y < x):
        return z
    else:
        return tak(tak(x-1, y, z), tak(y-1, z, x), tak(z-1, x, y))

for i in range(4):
  t = time.clock()
  r = tak(24, 14, 6)
  t2 = time.clock()
  print "Result: ", r, "Time: ", (t2 - t)

dumb theme
And finally the IronLisp code (see that macro!!!):

(using System.Diagnostics)

(= tak (fn (x y z)
  (if (not (< y x))
    z
    (tak (tak (- x 1) y z)
         (tak (- y 1) z x)
         (tak (- z 1) x y) ))))

(= time (macro (x)
  `(do
      (= sw (startnew stopwatch))
      ,x
      (= ts (elapsed sw))
      (cwl "time: {0:f3}ms" (totalmilliseconds ts)))))

(to i 4
  (time (tak 24 14 6)))