Date: 2015-07-10
- add
setup_requires
to setup.py
Date: 2015-07-10
- add
arrays.local
package to setup.py
Date: 2015-06-22
New Features:
- arrays can be distributed along multiple axes
- new distributed array:
pyDive.arrays.gpu_ndarray
. Distributed version of a pycuda array with extra support for non-contiguous memory (needed for working with sliced arrays). gather()
is called implicitly when the user wants to access an attribute, e.g. datapointer, which only the local array has.- implement indexing by bitmask
Bug Fixes:
- integer indices were not wrapped in
__getitem__()
- fix exception in picongpu.py if dataset has no attributes
Misc-
distaxes
parameter defaults to 'all'- all local array classes are located in pyDive/arrays/local
- rename "arrayOfStructs" module to "structured"
Date: 2015-03-29
New Features:
- adios support
- implement
pyDive.fragment
- distributed array classes are auto-generated from local array class.
Misc:
- restructure project files
- optimize
pyDive.ndarray
Date: 2014-11-06
- add MANIFEST.in
Date: 2014-11-06
- fix setup.py
Date: 2014-11-05
Development status: beta