Hi all, I am working on parallelizing a symbolic matrix inversion algorithm. I have written a sequential matlab code that uses syms variables to hold the symbolic expressions(matrices to be precise) and in the end computes the inverse by substituting real values to the symbols to get a numerical expression. Do the GPUs support such symbolic computations to be run in parallel? Are there architectures that are specifically suited to parallel symbolic computations? Do let me know your thoughts and any resources you might know.
Thanks,
Suhas.