scripting languages vs statically compiled ones

[Andrew Durdin]

Concur. A while back I thought I could speed up a VB app that had
a lot of bit-twiddle operations by writting a C-based DLL for it.

Turns out it ran slower that way!

Having used the VC++ debugger to step through compiled VB code, I
can attest to it being compiled down to native. Ratio is much higher
than with, say C or C++... I'd WAG it at about 10:1 or slightly more.

On the other hand, I had to do base64 decoding of a string in a VB
app; I got a very significant speed increase by rewriting that portion
with a C DLL. So YMMV I guess

 

[Steve Holden]

On the other hand, I had to do base64 decoding of a string in a VB
app; I got a very significant speed increase by rewriting that portion
with a C DLL. So YMMV I guess

The difference is probably that bit-=twiddling operations require very
little cmpute per operation, so in those cases the call overhead is
likely to dominate.

regards
Steve

 

[Isaac Gouy]

(e-mail address removed) (Torben Ægidius Mogensen) wrote in message

You could use the "The Great Computer Language Shootout" at
http://shootout.alioth.debian.org/craps.php . It may give the
impression that you can't trust the numbers it gives, but I believe
this is mainly because it is more honest than the majority of other
language comparisons.

Thank you! That's very kind - some folk seem to assume it's because
the authors are plain stupid ;-)

However, the websites been rearranged over the last month, so it would
be better to start from one of the main pages:

http://shootout.alioth.debian.org/
http://shootout.alioth.debian.org/great/
http://shootout.alioth.debian.org/sandbox/

The main problem with this test is that the benchmark programs are fairly
small, so the results may not be representative of large programs.

We'd quite like to include larger programs - but would people
volunteer their time to write larger programs in all those languages?