Ok, new tiny-blocks tests:

I tested:

1) Inserting 100.000 rows into two tables.
2) Select 100.000 rows from a table.

First of all, running with background priority image isn't freezing, 
and FFI calls are not perceived. This is important because this allows 
me to still render seaside pages (for example) for another user, when 
there is one user making a big query.
Second, I followed rows insertion (simple "select count(*) from bla1" 
in psql tool) and the insertions is happening in parallel.

Of course FFI calls are still happening, and there are many 
"micro-blocks", so the total time expended in each process is equal to 
the addition of all FFI blocks + squeak calculations. But is important 
to note that (2 parallel queries time) < (query1 time + query2 time)...

I'm going to put all this benckmarks in the wiki page (and the sources 
in squeaksource)

Cheers,
Esteban


On 2008-08-20 17:17:13 -0300, Esteban Lorenzano <estebanlm at gmail.com> said:

> He, I'm testing it now... in a while I upload results :)
> 
> On 2008-08-20 17:13:31 -0300, "Ramon Leon" <ramon.leon at allresnet.com> said:
> 
>>> 
>>> [ self executeAndRetreiveAVeryLongQuery ] forkAt: self
>>> userBackgroundPriority.
>>> 
>>> And it works perfect!
>>> No freeze is observed :)
>> 
>> But will this work...
>> 
>> [ self executeAndRetreiveAVeryLongQuery ] forkAt: self
>> userBackgroundPriority.
>> 
>> [ self executeAndRetreiveAShortQuery ] forkAt: self userBackgroundPriority.
>> 
>> And the short query finish first even though the longer one started first?
>> 
>> Ramon Leon
>> http://onsmalltalk.com