Re: Help Data-Types

From: Paul Vernon <paul.vernon_at_ukk.ibmm.comm>
Date: Fri, 23 Aug 2002 11:28:57 +0100
Message-ID: <ak52v0$2chu$2_at_sp15at20.hursley.ibm.com>


> Please switch off the HTML - it is completely illegible on my
newsreader

Ah well, you shouldn't be using a reader that supports HTML :-). But seriously, I apologise, I think there is a bug in my newsreader that occasionally makes it send HTML (and at point size 2 !!!) even though all my options say plain text only. I would change newsreaders, but I'm too stubborn, sorry.

>In this time 2GHz processor can do additional
> 2 * 10^9 * 11 / ( 2 ^ 20 * 12 ) = 1748 cycles

>Still, I don't know the answer on the following questions:
>In this extra amount of CPU cycles enough to do a funciton like
>F.IPV4DSPLY?
In principle way more than enough. I'd guess a good assembly coder could do that function in heck I don't know, 4 cycles maybe less, depending on the instruction set and parallelism of the CPU. Lets see, take each byte as an offset to 256 row array storing the 3 ASCII characters and write the result with ASCII '.'s to an area of memory. How many cycles is that on a P3?

> What is min percentage of cycles eaten by OS and database engine?
> In what form are user defined functions in SQL executed? I mean what
> is the avg ratio beetween basic functions and operators in SQL and
> processor cycles (to what is it compiled to? what's the overhead?)?

I believe a lot of SQL get compiled to some kind of byte code but I'd really not know what the overhead is. Put it this way, I've never seen any

slow down in using a complex function compared to a straight select.

but anyway, I still claim that querying 2Tbs is faster than querying 4Tb

Regards
Paul Vernon
Business Intelligence, IBM Global Services Received on Fri Aug 23 2002 - 12:28:57 CEST

Original text of this message