Available with Geostatistical Analyst license.
Several steps in the calculation of geostatistical layers take advantage of the increased performance available in systems that use multiple CPUs (or multi-core CPUs). Parallel processing is used for raster creation in all interpolation methods, and for the model calculation in all kriging, inverse distance weighting, and local polynomial models created within the Geostatistical Wizard. Contour rendering, cross-validation, validation, and prediction functions for geostatistical layers also make use of multiple CPUs.
Additionally, the following tools in the Geostatistical Analyst toolbox take advantage of multiple CPUs:
- Areal Interpolation Layer To Polygons
- Densify Sampling Network
- Diffusion Interpolation With Barriers
- Empirical Bayesian Kriging
- GA Layer To Contour
- GA Layer To Grid
- GA Layer To Points
- Gaussian Geostatistical Simulations (for conditional simulation)
- Moving Window Kriging
Increased processing speed is most noticeable in complex models (for example, involving data transformations and large neighborhoods) and in repeated use of the tools (as in models and scripts).
The Geostatistical Wizard and tools mentioned above will automatically use up to 16 threads/CPUs. This number can be modified by adding GA_SMP_MAX_THREADS as a system variable to the Windows Environment Variables. The variable's value sets the upper limit for the number of threads/CPUs Geostatistical Analyst processes can use.