Open doofin opened 6 years ago
Are there plans for support gpu backend which is essential for machine learning? And how to contribute
I have heard of no such plans. ATM hmatrix uses blas and lapack. You'd have to look to see if there were versions of these that support GPUs.
An alternative would be to implement an accelerate backend for hmatrix.
Are there plans for support gpu backend which is essential for machine learning? And how to contribute