api_status.md

Array API Coverage Implementation Status

Cubed supports version 2022.12 of the Python array API standard, with a few exceptions noted below. The Fourier transform functions are not supported.

Support for version 2023.12 is tracked in Cubed issue #438.

This table shows which parts of the the Array API have been implemented in Cubed, and which ones are missing. The version column shows the version when the feature was added to the standard, for version 2022.12 or later.

Category	Object/Function	Implemented	Version	Notes
Array object	Arithmetic Ops	✅
	Array Ops	✅
	Bitwise Ops	✅
	Comparison Ops	✅
	In-place Ops	❌		Arrays are immutable
	Reflected Ops	✅
	Attributes	✅
	Methods	✅		Not device methods
Constants	e, inf, ...	✅
Creation Functions	arange	✅
	asarray	✅
	empty	✅
	empty_like	✅
	eye	✅
	from_dlpack	❌
	full	✅
	full_like	✅
	linspace	✅
	meshgrid	✅
	ones	✅
	ones_like	✅
	tril	✅
	triu	✅
	zeros	✅
	zeros_like	✅
Data Type Functions	astype	✅
	can_cast	✅
	finfo	✅
	iinfo	✅
	result_type	✅
Data Types	bool, int8, ...	✅
Elementwise Functions	add	✅		Example of a binary function
	negative	✅		Example of a unary function
	others	✅
Indexing	Single-axis	✅
	Multi-axis	✅
	Boolean array	❌		Shape is data dependent, #73
Indexing Functions	take	✅	2022.12
Inspection	capabilities	✅	2023.12
	default_device	✅	2023.12
	default_dtypes	✅	2023.12
	devices	✅	2023.12
	dtypes	✅	2023.12
Linear Algebra Functions	matmul	✅
	matrix_transpose	✅
	tensordot	✅
	vecdot	✅
Manipulation Functions	broadcast_arrays	✅
	broadcast_to	✅
	concat	✅
	expand_dims	✅
	flip	✅
	permute_dims	✅
	repeat	✅
	reshape	✅		Partial implementation
	roll	✅
	squeeze	✅
	stack	✅
	tile	❌	2023.12
	unstack	✅	2023.12
Searching Functions	argmax	✅
	argmin	✅
	nonzero	❌		Shape is data dependent
	searchsorted	❌	2023.12
	where	✅
Set Functions	unique_all	❌		Shape is data dependent
	unique_counts	❌		Shape is data dependent
	unique_inverse	❌		Shape is data dependent
	unique_values	❌		Shape is data dependent
Sorting Functions	argsort	❌		Not in Dask
	sort	❌		Not in Dask
Statistical Functions	cumulative_sum	❌	2023.12	WIP #531
	max	✅
	mean	✅
	min	✅
	prod	✅
	std	✅
	sum	✅
	var	✅
Utility Functions	all	✅
	any	✅

Linear Algebra Extension

A few of the linear algebra extension functions are supported, as indicated in this table.

Category	Object/Function	Implemented	Version	Notes
Linear Algebra Functions	cholesky	❌
	cross	❌
	det	❌
	diagonal	❌
	eigh	❌
	eigvalsh	❌
	inv	❌
	matmul	✅
	matrix_norm	❌
	matrix_power	❌
	matrix_rank	❌
	matrix_transpose	✅
	outer	✅
	pinv	❌
	qr	✅
	slogdet	❌
	solve	❌
	svd	✅
	svdvals	✅
	tensordot	✅
	trace	❌
	vecdot	✅
	vectornorm	❌