Release Notes
Changed
- Improved the code performance by creating a global thread pool that is
shared across all functions that run in parallel. This avoids creating
a new thread pool for each function call, which was causing a performance
overhead. - Improved the performance of the
gaussian_smoothingfunction by using
a more efficient algorithm to compute the convolution of the input
data with a Gaussian kernel.