[geeklog-devel] Pear DB Overhead

Tony Bibbs tony at tonybibbs.com
Thu Feb 6 18:57:24 EST 2003

Well, the requirements for GL2 as they stand now are to support multiple 
  backends.  ADODB and PEAR::DB do this and, according to that link I 
just posted, they are fairly comparable.  You give up speed for 
flexibility in this case.  We could write our own lite DBMS abstraction 
layer but I don't see much value in reinventing the wheel.

Tom, I think many of your issues can be addressed in code.  For example, 
the new session library will reduce the 3 or 4 calls made to get data 
about a user and compress it into one call.  I think there are many 
instances where DB_count in the 1.3.x code is used to get a count for 
something that really isn't needed.  In fact our mantra as we develop 
GL2 should be "less is more".  Are you willing to stick with MySQL as a 
requirement for GL2?  We can bring that back up as a discussion point if 
you want.  Personally, I think pigeon holing ourselves to one DBMS is 
unwise if you want to put GL2 in the hands of as many users as possible. 
   I'd appreciate the input from you other developers here as I don't 
want to sound like a lone-renegade programmer out doing whatever he 
wants, when he wants.  There was concensus on this at one point and we 
tossed around these very things.

What in the design of GL2 do you see getting worse?  I couldn't disagree 
more with your view point here.


Tom Willett wrote:
> So with the enhancements its only 1/2 speed instead of 1/3 speed?
> My concern is that even now people complain about speed.  When they see that 
> it takes twice as long to render a page, they will really complain.  Much of 
> the speed issue now is because of uneeded and redundant db calls done in the 
> name of programming ease.  As I see the design of GL2, this will only get 
> worse.  A bunch of independent modules do not tend to share the information 
> they pulled from the db.
> Tom
> On Thu, 6 Feb 2003 16:09:32 -0600 (CST), Tony Bibbs wrote
>>Read this:
>>Sleep easier now?
>>On Thu, 6 Feb 2003, Tom Willett wrote:
>>>Was surfing around and say this rather disturbing bit of info about php 
> and 
>>>databases.  Well to put it simply PEAR DB appears to be about three 
> times 
>>>slower that mysql -- here are the benchmark times.
>>>        Average	 Overhead
>>>MySQL	 1.14	  -
>>>dbx      1.37     20% (index only)
>>>ADODB	 1.45	  27%
>>>dbx      1.53     34% (index/assoc/info)
>>>PhpLib	 1.60	  40%
>>>MDB      1.75	  54%
>>>PEAR DB  2.87	 152% (fetchInto)
>>>PEAR DB  3.15	 176% (fetchRow)
>>>M'base	 2.52	 296% (numeric cols)
>>>M'base	 4.77	 318% (assoc cols)
>>>and a link to the full article.
>>>Tom Willett
>>>tomw at pigstye.net
>>>geeklog-devel mailing list
>>>geeklog-devel at lists.geeklog.net
>>Tony Bibbs         | "I guess you have to remember that those who don't 
> |   
>>tony at tonybibbs.com | hunt or fish often see those of us who do as       |
>>                   | harmlessly strange and sort of amusing. When you   | 
>>                   | think about it, that might be a fair assessment."  | 
>>                   | --Unknown                                          |
>>geeklog-devel mailing list
>>geeklog-devel at lists.geeklog.net
> --
> Tom Willett
> tomw at pigstye.net
> _______________________________________________
> geeklog-devel mailing list
> geeklog-devel at lists.geeklog.net
> http://lists.geeklog.net/listinfo/geeklog-devel

|Tony Bibbs         |[R]egardless of what you may think of our penal   |
|tony at tonybibbs.com |system, the fact is that every man in jail is one |
|                   |less potential fisherman to clutter up your       |
|                   |favorite pool or pond. --Ed Zern                  | 


More information about the geeklog-devel mailing list