March 2024 Issue Index
The quest for speed
Maptek continually implements new approaches to processing to enhance the performance of software applications.
Multi-threaded grade estimation
Vulcan provides the capability to parallelise computational intense processes including grade estimation of numeric variables into 3D block models. The functionality has been available for many years and as processors get faster, Maptek has tapped into this power to benefit users.
Central processing unit (CPU) cores are accessed to run multiple processes simultaneously, with the results merged seamlessly into the final model. It is not uncommon to have 20+ cores in current CPUs, each potentially available for processing. Multi-threading capability can be accessed interactively through the Vulcan interface or automated through workflows and/or scripts.
A relatively new Run Estimation Multi option parallelises the already multi-threaded system. Computation times for a large number of block estimation IDs can be reduced by an order of magnitude.
The faster you can convert data to a viable model, the faster you can deliver the updates for resource reporting and grade control.
Just-in-time compilation
Once a resource or grade control geological model has been generated, additional work is often required to compute geometallurgical, geomechanical or mine planning values.
This could involve adding a new variable to the block model and calculating the in-situ dollar value of a block based on a simple set of pricing and recovery factors. Other scenarios may incorporate complex calculations based on parameters impacting the entire mine extraction to metallurgical recovery system.
These block calculation scripts can be multi-threaded to the computer CPU cores as with the grade estimation multi-threading. For even better performance, just-in-time compilation provides a result two to three times faster.
Multi-threading can be enabled in Vulcan preferences.
Cloud processing
As Maptek develops for the future, many of the compute intensive tasks will use cloud processing to deliver results much more quickly. Existing applications, DomainMCF and Evolution already work with cloud hosted GPUs and CPUs.
Data and process security is of utmost importance when sourcing compute resources outside your organisation and our existing experience delivering commercial applications in this manner will provide confidence to all users.
Contact Maptek technical support to see how you can speed up your work.
- A Run Estimation Multi option parallelises the already multi-threaded system, reducing computation times by an order of magnitude
- Block calculation scripts can be multi-threaded to your computer CPU cores, with just-in-time compilation providing a result two to three times faster
- Data and process security is of utmost importance when compute intensive tasks use cloud processing to deliver faster results