Searching for some terms with Google I came across this strange CUDA project proposal. On first brief skimming the text I thought this made no sense at all to me. Politics and GPUs don’t mix.
[url=“http://www.seas.upenn.edu/~cis665/projects2008/ProjectPropsalStocker.pdf”]http://www.seas.upenn.edu/~cis665/projects...psalStocker.pdf[/url]
I wonder if this was ever implemented in hardware ;-)
Christian
The concept of “Support Vector Machine” which is at the core of this proposal has been used for many years for machine learning, which in turn uses regression, search and classification techniques. For short, SVM - is or can be used on compute clusters in a parallel fashion.
Politics - is just a subject matter here. You would be surprised how far “computational politics” had come in the past few years - and I do not mean “rigging voting machines here” -):
Machine learning is a huge field, but not to digress too much here. Yes, perfect idea for GPUs. The amount of damage you could unleash on regression testing, searching, decision making using a few hundred CUDA processors is significant.
joseph