| Re: [eigen] ThreadSafety of Eigen? |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] ThreadSafety of Eigen?
- From: Benoit Jacob <jacob.benoit.1@xxxxxxxxx>
- Date: Wed, 31 Mar 2010 07:46:50 -0400
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:received:message-id:subject:from:to:content-type :content-transfer-encoding; bh=ojWSC2ScvM0XzfjiY+959GC54oat63viiR+/c1lSR+k=; b=tIigOi0ydS6J1ehmZ5MQYwSbFkaQZUhMfiLL5fzgh9fT4lLbJT2fhJtRceF1Muf3Jh WicssTkzIdMlUG4Z4ef3WBko+S3c0dagXqJNRVJrdxs+p5QH3V+vNkZvG474hYdle+qq ONBmrOeyaQP2SMofI3SnXZIp0yPAVkj/vz2b8=
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type:content-transfer-encoding; b=LHWTogZQSZ5whReJ0aujbQTEYDmPTFIZvQq59pGFbTFfTPsLG34upieg+MPKj/Vt0E IkjR1jCFCCfdMhFNDT9BBNpyQLB7HYtWVkEeXr77BQovvWnqedk1u7OUfPNU9RCqFnVe SCLyPCvlRD8/sa8I4/yc79aeEYqVOD/XKJLXE=
Hi,
Eigen is state-less, hence completely thread-safe in the sense that if
you have two matrices m1 and m2, then you can have thread1 working on
m1 and thread2 working on m2 without any interference. Also, all
functions/methods are reentrant (no static variables...).
But if you let 2 threads work on the same matrix, Eigen by itself
doesn't do anything (like locking) to prevent bad stuff from
happening.
Benoit
2010/3/31 Schmidt, Michael <Michael.Schmidt@xxxxxxxxxxxxxxxxxxx>:
> Hi,
>
>
>
> I am working at a research institute in the field of medical image analysis
> and would like to use the Eigen library for some of my image processing
> algorithms for the calculation of eigenvalues- and vectors of hessian
> matrices of 3D medical images. Can you give me some hints on the
> thread-safety of the Eigen-library in general or the use of simple
> Matrix-types (e.g. Matrix3f) in conjunction with the SelfAdjointEigenSolver
> in special? I know I could try to investigate it on my own, but then again
> there are so many possible pitfalls in threadsafe development that I wanted
> to ask here first.
>
>
>
> Best regards,
>
>
>
> Michael
>
>
>
>
>
> Dipl.-Inf. Michael Schmidt
>
> Fraunhofer MEVIS
> Institute for Medical Image Computing
> Universitaetsallee 29
> 28359 Bremen, Germany
>
> Tel.: + 49 421 218 2959
> Fax.: + 49 421 218 4236
>
> michael.schmidt@xxxxxxxxxxxxxxxxxxx
> www.mevis.fraunhofer.de
>
>