[Ocaml-biz] application level benchmarks

[Brandon J. Van Every]

Brandon J. Van Every wrote:
>
> I would comment that if people want to market OCaml using benchmarks,
> they would do well to consider the most relevant points of comparison
> between C++, Java, and C#, and design a benchmark for those.

Performance might also become a tough argument to make in the face of
the latest Intel C++ compiler.
http://www.open-mag.com/754088105111.shtml  GCC and MSVC are easy
targets to defeat, but the Intel compiler appears to be raising the
performance standard to a new bar.  I don't think C++ guys will ever
switch to OCaml if the performance isn't in the same ballpark.  I don't
know that marketing can address such an issue... it may be down to what
the OCaml implementors choose to work on.

A 'hedge' might be to investigate the Intel C++ compiler's
application-level performance, rather than simple synthetic benchmarks
where presumably it can kick the most ass.  How do OCaml and Intel
compare at the application level?  Of course, one would need to devise a
fair test application.

Another 'hedge' would be to worry mainly about beating Java and C#.  At
present, these are much easier to go after.  One could do that with
either synthetic or application level benchmarks.  Again, in the latter
case one would need to design a 'fair' application.

When using benchmarks for marketing, there is of course the strategic
risk that C++, Java, or C# could improve their performance and the OCaml
implementation doesn't keep up.  It's one of the ways that windows of
marketing opportunity can close.  So I suppose OCaml shouldn't pitch its
tent solely upon the performance landscape.

I'm enjoying talking to myself, BTW.  It's quite peaceful.  Input any
time you feel like it.


Cheers,                     www.indiegamedesign.com
Brandon Van Every           Seattle, WA

"The pioneer is the one with the arrows in his back."
                          - anonymous entrepreneur