Re: Ada vs SQLite3 benchmark

["Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de>]

On Thu, 18 Sep 2014 15:41:25 +0200, Jacob Sparre Andersen wrote:

> Dmitry A. Kazakov wrote:
> 
>> I posted benchmark of Ada persistent B-tree vs. SQLite3 at Ada
>> Programming blog:
>>
>> http://ada-programming.blogspot.de/2014/09
> 
> Nice write-up.
> 
> One thing which I miss is some sort of error-bars on your measurements.

Yes, but I could not provide them. In the sense that of course I could
measure single operation duration within the loop, and accumulate minimum
and maximum times. But the maximum time would depend on the OS process
switching and the time slice etc. To remove these would be too complicated,
so I used only averaged time.

> One argument for using SQLite, which you don't cover, is the command
> line tool for inspecting a SQLite database.  Have you considered how one
> could generate something like that for your persisten B-tree library?

No, because it is not meant to be relational. You could have as many
B-trees allocated there as you wished plus allocated objects out of any
tree, as you would do by Ada "new T". So basically it lacks any predefined
structure. It is just a memory pool.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de