Although WQS is multi threaded and seems to be able to handle all the CPU cores we can throw at it, we are finding that run times are prohibitively long for complex RBD models utilising resources and spares optimisation. We'd like to request one (or both!) of the two options below to help decrease processing time.
- Support for GPU processing (NVidia Tesla GPUs). This would allow for significantly more parallel calculations to be performed at once as the GPU contains thousands of cores.
- Support for calculations across a cluster of computers. Adding more CPU cores to a single computer is only economical up to a point. Having the ability to process across a cluster of computers will allow for additional parallel processing without the need to purchase exotic hardware.