Subject Re: [ib-support] IB7 vs. Firebird 1.0x features
Author Phil Shrimpton
On Sunday 16 February 2003 22:09, you wrote:

Hi,

> "In prior versions of IB once a SQL statement begin
> execution on a single CPU machine other pending tasks would
> not gain access to the CPU until the executing statement
> finished."
>
> Does this statement apply to Firebird 1.0x as well?

Only in SS, CS version is not affected.

> If it does, how serious a problem is this in real world
> situations?

If somebody is doing a long, intestive 'query' (e.g. for a report) it can
slow down other users. In reality, in a well designed database, not a major
problem, but something that does need addressing.

> "InterBase 7 does not support decomposing a single SQL
> statement into multiple execution threads that can run
> simultaneously on multiple processors."
>
> Any plans to support this in future versions of Firebird?

SMP support is planned, but when and in what form I can't answer. CS has
alsways supported multi-cpu's and the 'classic' version build has been made
avaiable for Windows for a current solution

> Now install InterBase 7 and repeat the same test. You will
> see both CPU's being used at the same time."
>
> Does this imply that this is not yet the case with Firebird
> 1.0x? I thought one could already expect this using Linux
> Firebird 1.0x SS/CS and NT Firebird NT CS?

Again its only SS that has the 'issue', CS is fine.

> "Temporary Tables for Monitoring the InterBase Server"
>
> Will Firebird 1.5 or later attempt to be compatible with these?

I like the funtionality, but the jury is out on the implementation. As for
FB plands to implement them, ther is probably a feature request or twelve,
but as with any open source project, new feature depend on the time and
effort of people working in their free time. Of course, a sizable
sponsership to the "Firebird Foundation" might help <g>

As for attempting to be 'compatible' with IB7; If/when this functionality is
implement I would argue that the developers do in the way they best see fit,
rather than trying to fit it around another databases implementation

Phil

--
Linux 2.4.4-4GB
8:31pm up 34 days, 2:20, 1 user, load average: 0.27, 0.08, 0.02