Re: Help Data-Types
Date: Tue, 20 Aug 2002 17:51:09 +0100
Message-ID: <ajts7d$3qpi$1_at_sp15at20.hursley.ibm.com>
>>My betting is that a CPU can do bit shifts quicker that getting a 11
more
>>bytes from RAM/disk
> Seems reasonable - but is it really so?
Of course it all depends where that data is. The outer edge of a disk when
the head is at the inner? Well a 32 way SMP clocked at say 2GHz with 4
execution units in each CPU could theoretically notchup say 4 * 2*10^9 *
32 / .2 instructions in the 200 milliseconds it takes to retrieve that
extra data. Even getting stuff from Level2, 3 & 4 RAM takes time.
Plus you only need to do the bit shifts/arithmetic/string ops when
displaying results to users. Joins, sorts etc can use the 'internal
format'.
On the other hand with a M/F with lots of I/O and not so much CPU, biasing
internal representations away from those with compute heavy requirements
can be a sensiable idea (e.g. store DATE TIMEs in BCD not num secs since
1971).
> I am asking since you did do a project with that representation and I
> guess that you did some testing
Nope, it's just common sense (or not so common apparently...) Think of it this way. Which is quicker, creating a summary table from 2 TBs of data or 4TBs? Answers on a postcard please.
>>With a DBMS that supports multiple possible representations (of
DataTypes)
>>one does not need to choose between different representation at the
>>logical level.
>I was under impression that by deciding between 4 properties and 1
>property of appropriate domains one chooses more than a physical
>representation.
>I thought that the process of determining weather IP is a scalar value
>property or non-scalar falls under logical modeling regardless of
>representation?
With a DBMS that supports multiple possible expressible databases (i.e. view/table interchangeability as long as they are information equivalent), one does not need to choose a primary representation at the logical level. IP is scalar or it is non-scalar depending on which expressible database a user happens to be looking at.
Rather a nice parallel with multiple possible representations I think.
Anyone out there got a DBMS that supports such things?!?
Regards
Paul Vernon
Business Intelligence, IBM Global Services
Received on Tue Aug 20 2002 - 18:51:09 CEST