[OpendTect_Developers] SeisTrcWriter Parallel write

[Kristofer Tingdahl]

Hi Fernando,


I would suggest you look at SeisSequentialWriter, defined in seiswriter.h.

You can add traces in random order, and they will be written in the correct
order. It works as follows:

1. You announce that a trace is coming.
2. Once the trace is computed, it is submitted with submitTrace and it will
be written.

Have a look in src/Velocity/velocityvolumeconversion.cc  for an example on
the usage. There we have one reader reading in traces, the computation is
done in parallel, and written in the correct sequence.

Hope that helps,


Kristofer Tingdahl


On 14 April 2014 15:39, Fernando Bordignon <fbordignon at gmail.com> wrote:

> Hi guys. This time I have queries about writing to a seis cube in
> parallel. Can I call the writer.put(trc); by many threads at the same time?
> This is not critical for me, because I can mutex the writing process.
>
> Another issue is bugging me the most. Can I call writer.put(trc); out of
> traces order? For instance, I use threads to get the tracein:
> rdr_->get(trcin_) inside doWork of ParallelTask. After that, I write the
> traceout by calling: trcout_ = trcin_; (assuming this will copy trace
> location), write at trcout with my algorithm and finally do a
> writer->put(trcout_);
>
> The calls to writer->put will not necessarily occur in the same order of
> rdr_->get. I got this doubts because I'm getting out of order traces at the
> output cube. This is expected behavior or some OD bug?
>
> Thanks a lot,
> Fernando.
>
> _____________________________________________________________
> OpendTect Developers mailing list Developers at opendtect.org
> http://lists.opendtect.org/mailman/listinfo/developers
>
> You receive this mail because you are listed on developers at opendtect.orgTo unsubscribe please go to
> http://lists.opendtect.org/mailman/options/developers . If you encounter
> any problems, please contact support at opendtect.org .
>



-- 
*Dr. Kristofer Tingdahl**Chief Executive Officer*
*________________________________*

*dGB Earth Sciences*Phone:+31 534315155Skype:dgbtingdahlE-mail:
kristofer.tingdahl at dgbes.comInternet:dgbes.com & opendtect.org
*________________________________*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opendtect.org/pipermail/developers/attachments/20140414/b38cca5f/attachment.html>