[rrd-developers] Re: Some minor changes proposal

[Tobias Oetiker]

Today you sent me mail regarding Re: [rrd-developers] Re: Some minor...:

*> > *> Tobi, you wrote about the "INT" operator being more general useful. 
*> > *> I do agree about that but I do think that if you only want operators
*> > *> that fall in that category you should mention it now. I'm sure that
*> > *> not only I will come up with other operators that are not-so-general :)
*> > 
*> > I have no general opinion on these issues yet ... but we might have to do
*> > some performance evaluations of rrd_graph to see if the CDEF language
*> > verbosity has a siginificant impact ... 
*> > 
*> 
*> Am I right in assuming that parsing a line is only done once?
*> (inside rrd_graph(), during "case GF_CDEF")
*> 
*> If so: performance should not suffer a lot since the operators are 
*> looked up from an array. I could be wrong -I'm no programmer- but
*> I think that "9,2,/,INT" will cost more than "9,2,DIV" even if DIV
*> would not be programmed in that first case.
*> 
*> I'm sure there are C-experts here who can do that test.

the parsing is only done once but there is a switch/case statement for
evaluating the expressions. This sitts at the most inner loop of the CDEF
function ... I am not sure what the performance implications with large
switch statements are ... guess this can be optimized beautifully by the
compiler ...

cheers
tobi
*> 

-- 
 ______    __   _
/_  __/_  / /  (_) Oetiker, Timelord & SysMgr @ EE-Dept ETH-Zurich
 / // _ \/ _ \/ / TEL: +41(0)1-6325286  FAX:...1517  ICQ: 10419518 
/_/ \.__/_.__/_/ oetiker at ee.ethz.ch http://ee-staff.ethz.ch/~oetiker

--
* To unsubscribe from the rrd-developers mailing list, send a message with the
  subject: unsubscribe to rrd-developers-request at list.ee.ethz.ch