strategic intelligence

Strategic Intelligence

The emergence of GPGPU technologies in the 2000s was preceded by fifteen years or so of exploratory research.  Those developments fitted in well with our mandate to extend the sensitivity analysis and Marquardt-Levenberg regularisation methods originally commercialised by CGER during the 1970s (that earlier generation of software was designed to run on VAX computers and attachable array processor accelerators).  The resulting interest in accelerators and GPGPUs was shared with our collaborator BHP.

We developed a demonstration system with the same general style of software architecture that came to be used in the new generation of GPGPU systems.  We sought to have this commercialised but ran into a turbulent period in Australian science policy where several attempts to downgrade the local status of computationally-intensive scientific and engineering software and computational mathematics were being successfully resisted, albeit at a cost of considerable investor uncertainty and delay.

Our contributions to strengthening risk perceptions in respect of computationally-intensive scientific and engineering software and computational mathematics were to collaborate with the Propagate project and develop assured forms of strategic intelligence covering advanced scientific and engineering computation.  That round of work grew from a local base of expertise in information and innovation economics.