| Re: [eigen] assembling a sparse matrix from multiple threads |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
On 2018-07-05 14:47, Andrea Bocci wrote:
Yes, unfortunately: if I understand correctly how it works, a sparse matrix
does not preallocate all its elements; they are allocated as the matrix is
filled, and this part that is not safe for concurrent operations.
You can actually pre-allocate a sparse matrix (using
SparseMatrix::reserve():
http://eigen.tuxfamily.org/dox/classEigen_1_1SparseMatrix.html#ac2b219eb36fbab0bfae535cfbfc74a76
Even then (i.e., if you can guarantee that sufficiently many elements
are reserved per inner vector), insertion is still not completely
thread-safe, because elements can get moved, and the `m_data.m_size`
member variable will be modified by different threads -- you may get
lucky here, but you should really not rely on that.
Christoph
--
Dr.-Ing. Christoph Hertzberg
Besuchsadresse der Nebengeschäftsstelle:
DFKI GmbH
Robotics Innovation Center
Robert-Hooke-Straße 5
28359 Bremen, Germany
Postadresse der Hauptgeschäftsstelle Standort Bremen:
DFKI GmbH
Robotics Innovation Center
Robert-Hooke-Straße 1
28359 Bremen, Germany
Tel.: +49 421 178 45-4021
Zentrale: +49 421 178 45-0
E-Mail: christoph.hertzberg@xxxxxxx
Weitere Informationen: http://www.dfki.de/robotik
-----------------------------------------------------------------------
Deutsches Forschungszentrum fuer Kuenstliche Intelligenz GmbH
Firmensitz: Trippstadter Straße 122, D-67663 Kaiserslautern
Geschaeftsfuehrung: Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster
(Vorsitzender) Dr. Walter Olthoff
Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes
Amtsgericht Kaiserslautern, HRB 2313
Sitz der Gesellschaft: Kaiserslautern (HRB 2313)
USt-Id.Nr.: DE 148646973
Steuernummer: 19/672/50006
-----------------------------------------------------------------------