v1.2.1

Includes bug fix #54

v1.2.0

Due to removal of deprecated real-to-real FFT with DFTI_CONJUGATE_EVEN_STORAGE=DFTI_COMPLEX_REAL and DFTI_PACKED_FORMAT=DFTI_PACK
from Intel(R) Math Kernel Library, reimplemented mkl_fft.rfft and mkl_fft.irfft to use real-to-complex functionality with subsequent
copying to rearange the transform as expected of mkl_fft.rfft, with the associated performance penalty. The use of the real-to-complex
transform improves multi-core utilization which may offset the performance loss incurred due to copying.

v1.1.0

• Added scipy.fft backend, see #42
• Fixed #46

v1.0.14

Fix for bug in irfftn_numpy for non-default axes.

Update to NumPy interface to adjust to changes to pocketfft in NumPy 1.17

v1.0.13

Fixed memory leak, #39.

v1.0.12

Fix for issue #37, as well as performance improvement for fftn and ifftn on real arrays for binaries built with Intel(R) C Compiler.

v1.0.11

Inlined multi-iterator routines to improve efficiency of complex ND transforms on real vectors.

v1.0.10

Fixes issues #29.

Adds Travis integration.

v1.0.9

This is rerelease of 1.0.7, which adds include of mkl.h in the header file to fix build break with clang on MacOS.

v1.0.8

This is re-release of 1.0.7 with a fix for gcc compilation error, missed due to lack of CI.