|
1 | | -Version 2022.1.0 (2022-01-XX) |
| 1 | +Version 2022.1 (2022-02-03) |
2 | 2 | ----------------------------- |
3 | 3 | * Added accuracy unit tests, which can be used systematically |
4 | 4 | using the 'pyvkfft-test' installed script |
5 | 5 | * An extensive testing is now made before official releases, |
6 | 6 | evaluating all type of transforms (c2c, r2c, dct, 1, 2 and 3D, |
7 | 7 | in and out-of-place, norm 0 and 1), different GPUs, both OpenCL |
8 | 8 | and CUDA, etc... Comparison is made against pyfftw, scipy or numpy. |
9 | | -* Update to VkFFT 1.2.20, with support for DCT types 1, 2, 3 and 4, |
| 9 | +* Update to VkFFT 1.2.21, with support for DCT types 1, 2, 3 and 4, |
10 | 10 | also fixing a number of issues (see closed issues at |
11 | | - https://github.com/DTolm/VkFFT/issues) |
| 11 | + https://github.com/DTolm/VkFFT/issues), and passing all tests |
| 12 | + on different GPUs (OpenCL and CUDA, AMD and nVidia) |
12 | 13 | * Raise a RuntimeError if the VkFFTApp initialisation or the |
13 | 14 | GPU kernel launch fails, with the corresponding VkFFT error. |
14 | 15 | * [BUG] Correct inverse FFT calculation using pyvkfft.fft.ifftn() |
|
0 commit comments