On 10/01/2013, at 2:37 PM, resolute wrote:
> HelloWorld test with python
> root@resolute-ubuntu:/usr/local/src# time flx hello.flx
> Hello World
>
> real 0m1.111s
> user 0m1.028s
> sys 0m0.052s
> root@resolute-ubuntu:/usr/local/src# time python he.py
> helloWOrld
>
> real 0m0.016s
> user 0m0.004s
> sys 0m0.008s
>
> Is that something wrong with my operate?
>
Run the Felix again. The first time it has to compile the code
from Felix to C++ and C++ to binary. The second time it
just checks time stamps and runs the binary.
Also hello world is not a realistic test, Python may run this faster
even after compilation to binary because the execution time
of the actual code, for Python, or Felix, is only a tiny fraction
of the run time: actually loading the code and starting up
may take longer.
~/felix>flx speed/ack/felix/test.flx
Ack(3,13): 65533
4.91068
is a more realistic test. That's using clang -O1.
HAHA OMG:
~/felix>flx -O2 speed/ack/felix/test.flx
Ack(3,13): 65533
7.56319
Clang is generating SLOWER code with -O2 than -O1!
here's C (unoptimised):
~/felix>gcc speed/ack/c/ack.c -o a.out
~/felix>time ./a.out
Ack(3,13): 65533
25.252633
real 0m25.295s
user 0m25.203s
sys 0m0.053s
gcc with -O2 does better:
~/felix>gcc speed/ack/c/ack.c -O2 -o a.out
~/felix>time ./a.out
Ack(3,13): 65533
8.356915
real 0m8.372s
user 0m8.345s
sys 0m0.015s
And clang 3,.3 with -O2;
~/felix>clang speed/ack/c/ack.c -O2 -o a.out
~/felix>time ./a.out
Ack(3,13): 65533
7.522444
real 0m7.531s
user 0m7.512s
sys 0m0.014s
So Felix TRASHES gcc and clang.
Here's the Python:
def ack(x,y):
if x == 0:
return y + 1
elif y == 0:
return ack(x - 1, 1)
else:
return ack(x - 1, ack(x, y - 1))
n = 13
v = ack(3,n)
print("Ack(3,%d): %d" % (n, v))
And here's the output:
RuntimeError: maximum recursion depth exceeded in comparison
Oops. Felix is infinitely faster.
Here with n=6, the maximum I can do on my box:
~/felix>time python3 ack.py
Ack(3,6): 509
real 0m0.137s
user 0m0.113s
sys 0m0.020s
And Felix:
~/felix>time flx ack
Ack(3,6): 509
0.000298023
real 0m6.782s
user 0m6.521s
sys 0m0.236s
~/felix>time flx ack
Ack(3,6): 509
0.000296116
real 0m0.038s
user 0m0.015s
sys 0m0.017s
The first time is slower because it takes 6 seconds to compile.
This isn't a realistic test. Python simply cannot do this.
The performance compared to C here is no accident,
but I have seen this vary a huge amount.
--
john skaller
skal...@users.sourceforge.net
http://felix-lang.org
------------------------------------------------------------------------------
Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
MVPs and experts. ON SALE this month only -- learn more at:
http://p.sf.net/sfu/learnmore_122712
_______________________________________________
Felix-language mailing list
Felix-language@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/felix-language
