Re: Euphoria Compilers Available Soon
>You don't have to repeat yourself... that's why I added the second
>question. If at one point I had a Euphoria interpretter that was faster
>than the one from RDS, then I would have saved that code and shared my
>implementation with those who write programs with Euphoria. *Then* I would
>have proceeded with making a compiler (but I would still save the code for
>this faster interpretter so that I could back up my claims). That's all...
>
>You've got to expect a little skepticism when you make claims that you
>can't immediately back up. (Sorry for being a skeptic, I will not make any
>further comments until you've released something that I can comment on.)
>
>-- Brian
Yeah man look.
The *project* started out as a faster Euphoria interpretter.
I never said I completed a full working Euphoria Interpretter that was
faster than RDS's,if I had,I wouldn't be creating a compiler at all.
I did test sequence appending/prepending with the source I had, and it took
50 milliseconds to append 5000 sequences to another. When doing the same
with RDS's Euphoria,I got the exact results. My method of appending was
calling realloc 5000 times,and a for loop that initialised the sequence from
scratch over and over again. That was bad code,very bad code,yet still as
fast as RDS's sequence appending.
What RDS should do to be considered producing a valid software creation
medium,is to get rid of Watcom and get Visual C++ for Win32 and DJGPP for
DOS. Watcom is slower than DJGPP and VC++ 6.0 is currently producing the
fastest code available on any PC platform.
Euphoria can be sped up by 30% just by switching compilers. If they can't
afford VC++, they can feel free to ask me for a copy of Enterprise Edition
5,I have a Microsoft license to install that software on 10 computers.
Mike The Spike
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
|
Not Categorized, Please Help
|
|
