OK, that may limit my work somewhat...In worst case I can have sparse matrices that are 4000x4000 entries in size- that is kind of huge, all things considered. When do you expect to optimize for sparse data and/or leverage the native provider for it?
↧