multi threading issues ...
Thomas Binder
binder at iue.tuwien.ac.at
Fri Mar 24 12:20:16 EST 2000
Hi,
I am using GALIB in a TCAD optimization framework which is capable of starting
several simulations on a cluster of workstations.
Therefore I am using a multi-threaded population evaluation mechanism; every
evaluation of a certain population is carried out in a separate thread. Each
thread queues a simulation run; the results from the simulations are then
collected from the main thread via stdin and passed to the (blocking) threads
via `userData'.
This mechanism worked fine until I started to change the `nBestGenomes'
parameter of the genetic algorithm. The program now stops with an exit status
of 139? as soon as the evaluation of the first population is finished.
I wrote a little test program which invokes the very same genetic algorithm
with completely the same settings but without multi-threaded population
evaluation and of course everything works fine. The only difference in the
little test program (apart from single-threaded evaluation) is that I use a
simple objective function (instead of the simulation run).
At this point I have a few questions:
Is the data in the multi-threaded (GA-)optimizer consistent even if I don't
touch `nBestGenomes'? If not, is there an easy workaround?
Has anybody else similar experiences/problems with thread-safety?
Any response would be appreciated
Thx,
Tom
--
"Computers are like air conditioners - they stop working properly when you open
Windows"
\\\|///
\\ ~ ~ //
( @ @ )
/------------------------oOOo-(_)-oOOo-----------------------------------------\
| Thomas Binder | |
| Institute for Microelectronics | phone: ++43/1/58801-36036 |
| Technical University Vienna | |
| Gusshausstrasse 27-29 / E360 | fax : ++43/1/58801-36099 |
| A-1040 Vienna | |
| A U S T R I A | email: binder at iue.tuwien.ac.at |
\---------------------------------Oooo.----------------------------------------/
.oooO ( )
( ) ) /
\ ( (_/
\_)
More information about the galib
mailing list