nums.core.compute.compute_interface module

class nums.core.compute.compute_interface.ComputeImp

Bases: object

class nums.core.compute.compute_interface.ComputeInterface

Bases: object

advanced_assign_block_along_axis(dst_arr, src_arr, ss, axis, dst_coord, src_coord, syskwargs)

advanced_select_block_along_axis(dst_arr, src_arr, ss, dst_axis, src_axis, dst_coord, src_coord, syskwargs)

arange(start, stop, step, dtype, syskwargs)

arg_op(op_name, arr, block_slice, other_argoptima, other_optima, syskwargs)

array_compare(func_name, a, b, args, syskwargs)

astype(arr, dtype_str, syskwargs)

bop(op, a1, a2, a1_T, a2_T, axes, syskwargs)

bop_reduce(op, a1, a2, a1_T, a2_T, syskwargs)

cholesky(arr, syskwargs)

create_block(*src_arrs, src_params, dst_params, dst_shape, dst_shape_bc, syskwargs)

diag(arr, syskwargs)

identity(value)

inv(arr, syskwargs)

logical_and(*bool_list, syskwargs)

map_uop(op_name, arr, args, kwargs, syskwargs)

new_block(op_name, grid_entry, grid_meta, syskwargs)

percentiles_from_tdigest(q, *digests, syskwargs)

permutation(rng_params, size, syskwargs)

pivot_partition(arr, pivot, op, syskwargs)

qr(*arrays, mode='reduced', axis=None, syskwargs)

random_block(rng_params, rfunc_name, rfunc_args, shape, dtype, syskwargs)

reduce_axis(op_name, arr, axis, keepdims, transposed, syskwargs)

reshape(arr, shape, syskwargs)

select_median(arr, syskwargs)

shape_dtype(arr, syskwargs)

size(arr, syskwargs)

split(arr, indices_or_sections, axis, transposed, syskwargs)

sum_reduce(*arrs, syskwargs)

svd(arr, syskwargs)

swapaxes(arr, axis1, axis2, syskwargs)

tdigest_chunk(arr, syskwargs)

touch(arr, syskwargs)

“Touch” the given array. This returns nothing and can be used to wait for a computation without pulling data to the head node.

transpose(arr, syskwargs)

update_block(dst_arr, *src_arrs, src_params, dst_params, syskwargs)

update_block_by_index(dst_arr, src_arr, index_pairs, syskwargs)

weighted_median(*arr_and_weights, syskwargs)

where(arr, x, y, block_slice_tuples, syskwargs)

xlogy(arr_x, arr_y, syskwargs)

class nums.core.compute.compute_interface.RNGInterface

Bases: object

new_block_rng_params()