|
| 1 | +### v3.2.20151211 |
| 2 | +- Feature parity with ArrayFire 3.2 libs |
| 3 | + - New computer vision functions: `sift`, `gloh`, `homography` |
| 4 | + - New graphics functions: `plot3`, `surface` |
| 5 | + - Functions to load and save native images: `load_image_native`, `save_image_native` |
| 6 | + - Use `unified` backend when possible |
| 7 | + |
| 8 | +- Added missing functions |
| 9 | + - `eval`, `init`, `convolve2_separable`, `as_type` method |
| 10 | + - `cuda` backend specific functions |
| 11 | + - `opencl` backend specific functions |
| 12 | + - `timeit` function to benchmark arrayfire functions |
| 13 | + |
| 14 | +- Added new examples |
| 15 | + - getting_started: `intro`, `convolve` |
| 16 | + - benchmarks: `bench_blas`, `bench_fft` |
| 17 | + - financial: `monte_carlo_options`, `black_scholes`, `heston_model` |
| 18 | + - graphics: `fractal`, `histogram`, `plot3d`, `conway`, `surface` |
| 19 | + |
| 20 | +- Bug fixes |
| 21 | + - Fixed bug when array types were being reported incorrectly |
| 22 | + - Fixed various bugs in graphics functions |
| 23 | + |
| 24 | +### v3.1.20151111 |
| 25 | +- Feature parity with ArrayFire 3.1 libs |
| 26 | +- Ability to interop with other python libs |
| 27 | +- Ability to extract raw device pointers |
| 28 | +- Load and Save arrays from disk |
| 29 | +- Improved `__repr__` support |
| 30 | + |
| 31 | +### v3.0.20150914 |
| 32 | +- Feature parity with ArrayFire 3.0 libs |
| 33 | +- Ability to switch all backends |
| 34 | +- Supports both python2 and python3 |
0 commit comments