What is cephes in v_exp_cephes_f32?

General Questions Feedback & Feature Request
1

What is the difference with “float64 v_exp_f32(float64 x)” and “float64 v_exp_f32(float64 x)”?
Can we do exp(scalar) like c++? The exponential operation on Scalar?

2

we can’t do exp(scalar) since all the these functions are through vector pipeline. If you want to use exp(scalar), CPU (like numpy) should be the best way, there is no benefit to HPU since it is only one data entry.

For cephes, please refer to this and this