We are steadily working towards a new version of PCRaster that is better capable of using the current generation of hardware resources. The goal is to make future versions of PCRaster execute models faster. More information about this work can be found on the new High-performance modelling parent project page.
We are glad to announce the final release of PCRaster-4.1.0! We fixed several bugs in the software and documentation, added support for pickling PCRaster Python types, and changed the location of shared libraries on Windows. For more information please visit the download page.
We released a beta version of the upcoming PCRaster-4.1.0 release. The (user visible) changes are listed here:
PCRaster packages are available for:
- Linux 64bit
- Windows 64bit
We dropped support for Window 32 bit. If 32 bit Windows is still important for you, you can stick with the previous version of PCRaster (4.0.2).
Please give this beta a spin if you have the chance. Note that since PCRaster 4, multiple versions of PCRaster can be installed side by side. You never have to uninstall a previous version if you prefer not to. Just install each new version next to the older one(s) and update the PATH and PYTHONPATH environment variables. If things don’t work for you, then please let us know, and revert the environment variables to their previous settings.
In case no show-stopper bugs are found, we plan to release the final version of PCRaster-4.1.0 about two weeks from now.
In a previous post we mentioned Canopy as a Python distribution that can be used to develop PCRaster Python models. The PCRaster Python package is compatible with other Python distributions though. One user informed us that he liked using Anaconda, and at our institute we installed WinPython for our students.
We are happy to announce the final release of PCRaster-4.0.0! For more information, visit the PCRaster 4.0.0 download page.
It is possible that your PCRaster models are not executing as fast as the could on your Linux system. That can happen because the default memory allocation and de-allocation rules that are in effect on Linux are optimized for system wide efficiency, instead of raw performance. For more information about this, see this document.
You can tune the memory allocation and de-allocation logic using environment variables (note the trailing underscore):
export MALLOC_MMAP_MAX_=0 export MALLOC_TRIM_THRESHOLD_=-1
You may want to check if these values help you to squeeze a bit of extra performance out of your models.
I have released a second test version of the upcoming PCRaster-4.0.0 release. See the PCRaster 4.0.0 download page for more information. This version has a much faster numpy2pcr operation and we fixed a bug in the resample command. Unless an important issue is found with this release, the next release will be the final 4.0.0 release.
I have released a test version of the upcoming PCRaster-4.0.0 release. See the PCRaster 4.0.0 download page for more information.
As of now, all PCRaster sources are distributed under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. All source code is stored on SourceForge.
I uploaded installers for BlockAnalyst to the Sourceforge download page. Check the BlockAnalyst project page for a link.