clesperanto kernel convertion effort #16

StRigaud commented 2 years ago

Related to #16 Checked one are either under process or finished. Once processed remove _xd and _x

[x] absolute_2d_x.cl
[x] absolute_3d_x.cl
[x] add_image_and_scalar_2d_x.cl
[x] add_image_and_scalar_3d_x.cl
[x] add_images_weighted_2d_x.cl
[x] add_images_weighted_3d_x.cl
[x] affine_transform_2d_interpolate_x.cl
[x] affine_transform_2d_x.cl
[x] affine_transform_3d_interpolate_x.cl
[x] affine_transform_3d_x.cl
[ ] apply_vectorfield_2d_interpolate_x.cl
[ ] apply_vectorfield_2d_x.cl
[ ] apply_vectorfield_3d_interpolate_x.cl
[ ] apply_vectorfield_3d_x.cl
[x] arg_maximum_z_projection_x.cl
[x] binary_and_2d_x.cl
[x] binary_and_3d_x.cl
[x] binary_edge_detection_2d_x.cl
[x] binary_edge_detection_3d_x.cl
[x] binary_not_2d_x.cl
[x] binary_not_3d_x.cl
[x] binary_or_2d_x.cl
[x] binary_or_3d_x.cl
[x] binary_subtract_2d_x.cl
[x] binary_subtract_3d_x.cl
[x] binary_xor_2d_x.cl
[x] binary_xor_3d_x.cl
[x] block_enumerate_x.cl
[x] convolve_2d_x.cl
[x] convolve_3d_x.cl
[x] copy_2d_x.cl
[x] copy_3d_x.cl
[x] copy_slice_from_3d_x.cl
[x] copy_slice_to_3d_x.cl
[ ] count_non_zero_projection_3d_2d_x.cl
[ ] count_nonzero_pixels_slice_by_slice_sphere_3d_x.cl
[ ] count_nonzero_pixels_sphere_2d_x.cl
[ ] count_nonzero_voxels_sphere_3d_x.cl
[x] count_touching_neighbors_x.cl
[x] crop_2d_x.cl
[x] crop_3d_x.cl
[x] detect_label_edges_2d_x.cl
[x] detect_label_edges_3d_x.cl
[x] detect_maxima_2d_x.cl
[x] detect_maxima_3d_slice_by_slice_x.cl
[x] detect_maxima_3d_x.cl
[x] detect_minima_2d_x.cl
[x] detect_minima_3d_slice_by_slice_x.cl
[x] detect_minima_3d_x.cl
[x] dilate_box_2d_x.cl
[x] dilate_box_3d_x.cl
[x] dilate_box_slice_by_slice_3d_x.cl
[x] dilate_sphere_2d_x.cl
[x] dilate_sphere_3d_x.cl
[x] dilate_sphere_slice_by_slice_3d_x.cl
[x] divide_images_2d_x.cl
[x] divide_images_3d_x.cl
[x] downsample_2d_x.cl
[x] downsample_3d_x.cl
[ ] downsample_xy_by_half_median_3d_x.cl
[x] draw_box_2d_x.cl
[x] draw_box_3d_x.cl
[x] draw_sphere_2d_x.cl
[x] draw_sphere_3d_x.cl
[x] drawline_2d_x.cl
[x] drawline_3d_x.cl
[ ] entropy_2d_x.cl
[ ] entropy_3d_x.cl
[x] equal_2d_x.cl
[x] equal_3d_x.cl
[x] equal_constant_2d_x.cl
[x] equal_constant_3d_x.cl
[x] erode_box_2d_x.cl
[x] erode_box_3d_x.cl
[x] erode_box_slice_by_slice_3d_x.cl
[x] erode_sphere_2d_x.cl
[x] erode_sphere_3d_x.cl
[x] erode_sphere_slice_by_slice_3d_x.cl
[ ] excludeLabelsOnEdges_3d_x.cl
[x] exponential_2d_x.cl
[x] exponential_3d_x.cl
[x] flag_existing_labels_x.cl
[x] flip_2d_x.cl
[x] flip_3d_x.cl
[x] flood_fill_diamond_x.cl
[x] gaussian_blur_separable_2d_x.cl
[x] gaussian_blur_separable_3d_x.cl
[x] gradient_x_2d_x.cl
[x] gradient_x_3d_x.cl
[x] gradient_y_2d_x.cl
[x] gradient_y_3d_x.cl
[x] gradient_z_3d_x.cl
[x] greater_2d_x.cl
[x] greater_3d_x.cl
[x] greater_constant_2d_x.cl
[x] greater_constant_3d_x.cl
[x] greater_or_equal_2d_x.cl
[x] greater_or_equal_3d_x.cl
[x] greater_or_equal_constant_2d_x.cl
[x] greater_or_equal_constant_3d_x.cl
[x] histogram_2d_x.cl
[x] histogram_3d_x.cl
[ ] label_spots_in_x.cl
[x] labelled_spots_to_point_list_x.cl
[x] laplace_box_2d_x.cl
[x] laplace_box_3d_x.cl
[x] laplace_diamond_2d_x.cl
[x] laplace_diamond_3d_x.cl
[x] logarithm_2d_x.cl
[x] logarithm_3d_x.cl
[x] mask_2d_x.cl
[x] mask_3d_x.cl
[x] mask_label_2d_x.cl
[x] mask_label_3d_x.cl
[ ] mask_stack_with_plane_3d_x.cl
[ ] matrix_multiply_float_buffer.cl
[x] maximum_image_and_scalar_2d_x.cl
[x] maximum_image_and_scalar_3d_x.cl
[x] maximum_images_2d_x.cl
[x] maximum_images_3d_x.cl
[x] maximum_octagon_box_2d_x.cl
[x] maximum_octagon_box_3d_x.cl
[x] maximum_octagon_diamond_2d_x.cl
[x] maximum_octagon_diamond_3d_x.cl
[ ] maximum_of_masked_pixels_3d_2d_x.cl
[ ] maximum_projection_3d_2d_x.cl
[x] maximum_separable_2d_x.cl
[x] maximum_separable_3d_x.cl
[x] maximum_sphere_2d_x.cl
[x] maximum_sphere_3d_x.cl
[x] maximum_x_projection_x.cl
[x] maximum_y_projection_x.cl
[x] maximum_z_projection_x.cl
[x] mean_separable_2d_x.cl
[x] mean_separable_3d_x.cl
[x] mean_sphere_2d_x.cl
[x] mean_sphere_3d_x.cl
[x] mean_x_projection_x.cl
[x] mean_y_projection_x.cl
[x] mean_z_projection_x.cl
[ ] median_box_2d_x.cl
[ ] median_box_3d_x.cl
[ ] median_sphere_2d_x.cl
[ ] median_sphere_3d_x.cl
[ ] median_z_projection_x.cl
[x] minimum_image_and_scalar_2d_x.cl
[x] minimum_image_and_scalar_3d_x.cl
[x] minimum_images_2d_x.cl
[x] minimum_images_3d_x.cl
[x] minimum_octagon_box_2d_x.cl
[x] minimum_octagon_box_3d_x.cl
[x] minimum_octagon_diamond_2d_x.cl
[x] minimum_octagon_diamond_3d_x.cl
[ ] minimum_of_masked_pixels_3d_2d_x.cl
[ ] minimum_projection_3d_2d_x.cl
[x] minimum_separable_2d_x.cl
[x] minimum_separable_3d_x.cl
[x] minimum_sphere_2d_x.cl
[x] minimum_sphere_3d_x.cl
[x] minimum_x_projection_x.cl
[x] minimum_y_projection_x.cl
[x] minimum_z_projection_x.cl
[x] multiply_image_and_coordinate_2d_x.cl
[x] multiply_image_and_coordinate_3d_x.cl
[x] multiply_image_and_scalar_2d_x.cl
[x] multiply_image_and_scalar_3d_x.cl
[x] multiply_images_2d_x.cl
[x] multiply_images_3d_x.cl
[x] multiply_matrix_x.cl
[ ] multiply_stack_with_plane_3d_x.cl
[ ] multiply_stack_with_scalars_x.cl
[x] nonzero_maximum_box_2d_x.cl
[x] nonzero_maximum_box_3d_x.cl
[x] nonzero_maximum_diamond_2d_x.cl
[x] nonzero_maximum_diamond_3d_x.cl
[x] nonzero_minimum_box_2d_x.cl
[x] nonzero_minimum_box_3d_x.cl
[x] nonzero_minimum_diamond_2d_x.cl
[x] nonzero_minimum_diamond_3d_x.cl
[x] not_equal_2d_x.cl
[x] not_equal_3d_x.cl
[x] not_equal_constant_2d_x.cl
[x] not_equal_constant_3d_x.cl
[x] onlyzero_overwrite_maximum_box_2d_x.cl
[x] onlyzero_overwrite_maximum_box_3d_x.cl
[x] onlyzero_overwrite_maximum_diamond_2d_x.cl
[x] onlyzero_overwrite_maximum_diamond_3d_x.cl
[x] paste_2d_x.cl
[x] paste_3d_x.cl
[x] power_2d_x.cl
[x] power_3d_x.cl
[x] power_images_2d_x.cl
[x] power_images_3d_x.cl
[x] replace_intensities_x.cl
[x] replace_intensity_x.cl
[x] replace_pixels_if_zero_x.cl
[x] reslice_bottom_3d_x.cl
[x] reslice_left_3d_x.cl
[ ] reslice_radial_interpolate_2d_x.cl
[ ] reslice_radial_interpolate_3d_x.cl
[x] reslice_right_3d_x.cl
[x] reslice_top_3d_x.cl
[x] rotate_clockwise_2d_x.cl
[x] rotate_clockwise_3d_x.cl
[x] rotate_counter_clockwise_2d_x.cl
[x] rotate_counter_clockwise_3d_x.cl
[x] set_2d_x.cl
[x] set_3d_x.cl
[x] set_column_2d_x.cl
[x] set_column_3d_x.cl
[x] set_image_borders_2d_x.cl
[x] set_image_borders_3d_x.cl
[x] set_nonzero_pixels_to_pixelindex_x.cl
[x] set_plane_2d_x.cl
[x] set_plane_3d_x.cl
[x] set_ramp_x_2d_x.cl
[x] set_ramp_x_3d_x.cl
[x] set_ramp_y_2d_x.cl
[x] set_ramp_y_3d_x.cl
[x] set_ramp_z_2d_x.cl
[x] set_ramp_z_3d_x.cl
[x] set_row_2d_x.cl
[x] set_row_3d_x.cl
[x] set_where_x_equals_y_2d_x.cl
[x] set_where_x_equals_y_3d_x.cl
[x] set_where_x_greater_than_y_2d_x.cl
[x] set_where_x_greater_than_y_3d_x.cl
[x] set_where_x_smaller_than_y_2d_x.cl
[x] set_where_x_smaller_than_y_3d_x.cl
[x] smaller_2d_x.cl
[x] smaller_3d_x.cl
[x] smaller_constant_2d_x.cl
[x] smaller_constant_3d_x.cl
[x] smaller_or_equal_2d_x.cl
[x] smaller_or_equal_3d_x.cl
[x] smaller_or_equal_constant_2d_x.cl
[x] smaller_or_equal_constant_3d_x.cl
[x] sobel_2d_x.cl
[x] sobel_3d_x.cl
[ ] standard_deviation_z_projection_x.cl
[x] subtract_image_from_scalar_2d_x.cl
[x] subtract_image_from_scalar_3d_x.cl
[x] sum_reduction_x.cl
[x] sum_x_projection_x.cl
[x] sum_y_projection_x.cl
[x] sum_z_projection_x.cl
[x] transpose_xy_3d_x.cl
[x] transpose_xz_3d_x.cl
[x] transpose_yz_3d_x.cl
[x] undefined_to_zero_x.cl
[ ] variance_2d_x.cl
[ ] variance_masked_projection_x.cl
[ ] variance_projection_x.cl
[x] write_values_to_positions_2d_x.cl
[x] write_values_to_positions_3d_x.cl

StRigaud commented 2 years ago

@haesleinhuepf Can you remove the kernels that cannot be nd form that list and kernel which are not relevant for now. thanks!

haesleinhuepf commented 2 years ago

Hi @StRigaud ,

I just updated the list in your post above. I'm keeping two sub-lists here. The first should be done, but those are no filters, those are matrix operations and fancy stuff. The entries in the second list may be dropped:

Special but cool to have (no real 2D/3D filters)

[ ] generate_binary_overlap_matrix_2d_x.cl
[ ] generate_binary_overlap_matrix_3d_x.cl
[ ] generate_distance_matrix_x.cl
[ ] generate_touch_matrix_2d_x.cl
[ ] generate_touch_matrix_3d_x.cl
[ ] average_distance_of_n_far_off_distances_x.cl
[ ] average_distance_of_n_shortest_distances_x.cl
[ ] average_distance_of_touching_neighbors_x.cl
[ ] distance_matrix_to_mesh_x.cl
[ ] maximum_of_touching_neighbors_x.cl
[ ] mean_of_touching_neighbors_x.cl
[ ] median_of_touching_neighbors_x.cl
[ ] minimum_of_touching_neighbors_x.cl
[ ] n_shortest_distances_x.cl
[ ] n_shortest_points_x.cl

Potentially to drop entirely

[ ] maximum_z_projection_bounded_3d_2d_x.cl
[ ] maximum_slice_by_slice_sphere_3d_x.cl
[ ] mean_slice_by_slice_sphere_3d_x.cl
[ ] mean_z_projection_bounded_3d_2d_x.cl
[ ] median_slice_by_slice_box_3d_x.cl
[ ] median_slice_by_slice_sphere_3d_x.cl
[ ] minimum_distance_of_touching_neighbors_x.cl
[ ] minimum_slice_by_slice_sphere_3d_x.cl
[ ] minimum_z_projection_bounded_3d_2d_x.cl
[ ] minimum_z_projection_thresholded_bounded_3d_2d_x.cl
[ ] pointindexlist_to_mesh_3d_x.cl
[ ] standard_deviation_of_touching_neighbors_x.cl
[ ] touch_matrix_to_mesh_3d_x.cl
[ ] depth_projection_interpolation_x.cl
[ ] depth_projection_x.cl
[ ] distancemap_localPositiveMinimum_box_2d_x.cl
[ ] distancemap_localPositiveMinimum_box_3d_x.cl
[ ] distancemap_localPositiveMinimum_diamond_2d_x.cl
[ ] distancemap_localPositiveMinimum_diamond_3d_x.cl
[ ] exclude_labels_on_surface_2d_x.cl
[ ] exclude_labels_on_surface_3d_x.cl
[ ] exclude_labels_sub_surface_2d_x.cl
[ ] exclude_labels_sub_surface_3d_x.cl
[ ] piv.cl
[ ] watershed_binarize_labelmap_2d_x.cl
[ ] watershed_binarize_labelmap_3d_x.cl
[ ] watershed_detect_maxima_region_box_2d.cl
[ ] watershed_detect_maxima_region_box_3d.cl
[ ] watershed_eliminate_wrong_maxima_2d_x.cl
[ ] watershed_eliminate_wrong_maxima_3d_x.cl
[ ] watershed_local_maximum_2d_x.cl
[ ] watershed_local_maximum_3d_x.cl

StRigaud commented 2 years ago

Question on kernel to be dropped

slice_by_slice kernels: cons: They are redundant with classic operation applied at dimension n-1 in a loop so can be dropped pros: They avoid host code from user and speed up some development are they useful or not?

watershed: is it because the algorithm is not optimized? is there a way to make it relevant to keep in GPU?

haesleinhuepf commented 2 years ago

watershed: is it because the algorithm is not optimized? is there a way to make it relevant to keep in GPU?

It doesn't work well. See here:

https://github.com/clij/clij2/issues/18

slice_by_slice kernels: cons: They are redundant with classic operation applied at dimension n-1 in a loop so can be dropped pros: They avoid host code from user and speed up some development are they useful or not?

I think we can have the same functionality, e.g. by setting sigma_z=0, e.g. in the Gaussian blur. Plus, this would also work with X and Y, without having to maintain extra kernels

StRigaud commented 2 years ago

I think we can have the same functionality, e.g. by setting sigma_z=0, e.g. in the Gaussian blur. Plus, this would also work with X and Y, without having to maintain extra kernels

indeed, though i will need to update kernels already pushed. ToDo later then :smile:

StRigaud commented 2 years ago

Continuing on the slice_by_slice kernels:

do we plan on keeping the other slice_by_slice kernel in the top list (e.g. count_nonzero_pixels_slice_by_slice_sphere_3d_x.cl)?
do we plan on keeping the one I have already updated (e.g. erode_box, dilate_box, detect_maxima, etc.) or do we want the same behaviour (3d data but z=0 mean slice_by_slice process)?

StRigaud commented 2 years ago

From top post list

Raising the interest of keep the following kernels:

maximum_projection_3d_2d_x.cl :arrow_right: same as maximum_z_projection
minimum_projection_3d_2d_x.cl :arrow_right: same as minimum_z_projection

Raising interest of keeping, or renaming and extend to all axis (x,y,z) the following kernels:

mask_stack_with_plane_3d_x.cl :arrow_right: mask volume with binary plane along depth axis
multiply_stack_with_plane_3d_x.cl :arrow_right: multiply volume with plane along depth axis
multiply_stack_with_scalars_x.cl :arrow_right: multiply volume with vector along depth axis
maximum_of_masked_pixels_3d_2d_x.cl :arrow_right: masked max projection along depth axis
minimum_of_masked_pixels_3d_2d_x.cl :arrow_right: masked min projection along depth
count_non_zero_projection_3d_2d_x.cl :arrow_right: same as other (min,max,sum,etc.) projection kernel

:question: kernels:

excludeLabelsOnEdges_3d_x.cl :arrow_right: contains multiple kernel inside, should maybe split them?

There are other kernels left in the list, a quick discussion about them could be nice to see how to extend them to clEsperanto standards.

clEsperanto / clij-opencl-kernels

clesperanto kernel convertion effort #16 #17