User 677b9c22ff
15-11-2006 23:15:47
Hi,
as recently posted http://www.chemaxon.com/forum/ftopic2218.html several CPUs can be used during the calculation of terms, resulting in very fast queries.
If I enter the Lipinski rule of five from the query chemical terms filter as:
(mass() <= 500) && (logP() <= 5) && (donorCount() <= 5) && (acceptorCount() <= 10)
it will assign 4 additional threads (because 4 terms are evaluated) and if I have a 4 core CPU it will spead up accordingly (100% CPU used).
However if I add a new column and enter the same term it will also
assign 4 new threads however the insertion of the values into the DB field
takes almost double the time than in the first example shown. The bottleneck seems the insertion of the values into the DB. So if this problem could somehow be solved even complex terms wouldnt be a time problem (I am not talking about some, but millions of substances).
Tobias
as recently posted http://www.chemaxon.com/forum/ftopic2218.html several CPUs can be used during the calculation of terms, resulting in very fast queries.
If I enter the Lipinski rule of five from the query chemical terms filter as:
(mass() <= 500) && (logP() <= 5) && (donorCount() <= 5) && (acceptorCount() <= 10)
it will assign 4 additional threads (because 4 terms are evaluated) and if I have a 4 core CPU it will spead up accordingly (100% CPU used).
However if I add a new column and enter the same term it will also
assign 4 new threads however the insertion of the values into the DB field
takes almost double the time than in the first example shown. The bottleneck seems the insertion of the values into the DB. So if this problem could somehow be solved even complex terms wouldnt be a time problem (I am not talking about some, but millions of substances).
Tobias