On Wed, Dec 14, 2022 at 1:44 PM Scott Ransom <sransom@nrao.edu> wrote:

On 12/14/22 3:01 PM, Carl Kleffner wrote:

> Hi Scott,

>

> may I ask you which kind of vector / matrix operation in extended precision (np.longdouble) is

> supported in 'pint' ? It can't be backed by the underlying blas library as extended precision

> functions are not supported there.

>

> Carl

Hi Carl,

The vast majority of them are simply a vector of np.longdoubles (which are usually times or time

differences) which are operated on by regular Numpy math functions to compute new quantities. For

example, polynomial computations base on time, many trigonometric operations on the times for orbit

calculations, things like that.

We don't really do big longdouble matrix operations.

Sounds like a good place to start looking for improvements is in addition and multiplication.

Chuck