4 ms·
Is OpenBLAS faster for large matrices because of optimizations or asymptotically faster algorithms? https://stackoverflow.com/a/11421344 https://stackoverflow.c
by nyanpasu64 5y ago
Is OpenBLAS faster for large matrices because of optimizations or asymptotically faster algorithms? https://stackoverflow.com/a/11421344 https://stackoverflow.com/a/11421344 claims BLAS doesn't use Strassen's algorithm.
- jleahy 5y agoDue to optimisations similar to those in this article. The asymptotically faster algorithms are not useful in practice.
- quanto 5y agoStrassen's algorithm is not numerically stable and thus is not normally used in floating point operations. Usually, the performance differences between numerical libraries, especially for something as routine as matrix computation, are due to engineering implementations (unless explicitly stated otherwise).
- sampo 5y agoI've read that Strassen's algorithm, and fast matrix multiplication algorithms in general, accumulate too much numerical errors to be useful with single precision (f32) floats. That's why they were not in use before year 2000. With f64 it's better, and in the past 20 years there's been research on how to augment the fast algorithms to decrease the numerical errors. But those extra steps also make the fast algorithms slower. I don't know what the current situation is.