[Biophp-dev] mmCIF and Zend compiler

biophp-dev@bioinformatics.org biophp-dev@bioinformatics.org
Sat, 07 Jun 2003 9:59:07 PDT


> >>Mini-Me: There are criticisms of PHP being >>unsuitable for
> computationally-intensive tasks >>because it is an interpreted
> language.  Do you think >>the Zend compiler is the answer to this?
> Does it >>produce code whose speed is comparable to that >>produced
> by a C/C++ compiler?
> 
> >Nico: Yes, since it is code produces by a C >compiler.  
> 
> Ah, now we can get down to serious comp. bio. stuff.  Fro some time,
> I've shied away from writing the more  computationally-intensive
> algorithms/modules knowing that it's not one of PHP's strengths.  
> 
> But if what you're saying is true, then its "no holds barred, take no 
> prisoners" from now on. =) 
> 
> Oh, but first I'd have to raise the $ 2,000 to afford a copy of Zend. 
> *chuckle*  Or are they giving it away now? *fingers crossed*

You have always been free to write your own C-code and compile it in with 
php.  There is extensive documentation about this (I send the link in my
previous mail on the topic).  Free as in beer.  Again, unpracticle for
distribution since everyone who wants to use it will need to patch
his/her php version.  It is the Zend 'accelerator' that costs money, but
the free(albeit not Open Source) appears to be equally fast.


> >Nico: B.t.w.  there is a nice (and free) accelerator >for php
> available (www.php-accelerator.co.uk/).  It >caches (compiled) php
> code on disk and in RAM. I have >it installed on our server and it
> works flawlessly.
> 
> Thanks for that tidbit.  I'll check it out pronto.
> 
> Btw, can anyone take a look at my MySQL scripts and
> convert them to Pear or ADODB syntax?  You can find them here:
> 
>   http://genephp.sourceforge.net/mysql_dbscripts.html
> 
> While I've been using ADODB for some time (via Phakt), I've never
> really paid close attention to its (finer) syntax.  PEAR DB I 

I will. I will be using Adodb though.  Two reasons: 1. I have used that
code extensively and I am not going to use time to figure out PEAR DB.
2. Adodb is (claims to be) two times as fast as PEAR DB.


haven't 
> used much. 
> 
> I'm thinking of extending the I/O demo scripts (see URL below) to
> generate SQL scripts (INSERTs) for data that have been parsed. 
> 
>   http://genephp.sourceforge.net/io_set.html
> 

Good plan.  We should be able to use these tables to write seq objects to 
an SQL database and get them back out again.  I'll have a look.

Best,


Nico