Project comparisons#

Here is a list of NumPy APIs and corresponding cuPyNumeric implementations.

A dot in the cupynumeric column denotes that cuPyNumeric implementation is not provided yet. We welcome contributions for these functions.

NumPy vs cuPyNumeric APIs#

Module-Level#

NumPy	cupynumeric	single-GPU/CPU	multi-GPU/CPU
`numpy.DataSource`
`numpy.RankWarning`
`numpy.arange`	`cupynumeric.arange`	✓	✓
`numpy.array`	`cupynumeric.array`	✓	✓
`numpy.asanyarray`
`numpy.asarray`	`cupynumeric.asarray`	✓	✓
`numpy.asarray_chkfinite`
`numpy.ascontiguousarray`
`numpy.asfortranarray`
`numpy.asmatrix`
`numpy.bartlett`	`cupynumeric.bartlett`	✓	✓
`numpy.base_repr`
`numpy.binary_repr`
`numpy.blackman`	`cupynumeric.blackman`	✓	✓
`numpy.bmat`
`numpy.bool_`
`numpy.broadcast`
`numpy.broadcast_shapes`	`cupynumeric.broadcast_shapes`	✓	✓
`numpy.busdaycalendar`
`numpy.byte`
`numpy.byte_bounds`
`numpy.bytes_`
`numpy.cdouble`
`numpy.cfloat`
`numpy.character`
`numpy.chararray`
`numpy.clongdouble`
`numpy.clongfloat`
`numpy.complex128`
`numpy.complex256`
`numpy.complex64`
`numpy.complex_`
`numpy.complexfloating`
`numpy.csingle`
`numpy.datetime64`
`numpy.datetime_data`
`numpy.deprecate`
`numpy.deprecate_with_doc`
`numpy.diag_indices`	`cupynumeric.diag_indices`	✓	✓
`numpy.disp`
`numpy.double`
`numpy.dtype`
`numpy.empty`	`cupynumeric.empty`	✓	✓
`numpy.errstate`
`numpy.eye`	`cupynumeric.eye`	✓	✓
`numpy.find_common_type`	`cupynumeric.find_common_type`	❌	❌
`numpy.finfo`
`numpy.flatiter`
`numpy.flexible`
`numpy.float128`
`numpy.float16`
`numpy.float32`
`numpy.float64`
`numpy.float_`
`numpy.floating`
`numpy.format_float_positional`
`numpy.format_float_scientific`
`numpy.format_parser`
`numpy.from_dlpack`
`numpy.frombuffer`
`numpy.fromfile`
`numpy.fromfunction`
`numpy.fromiter`
`numpy.frompyfunc`
`numpy.fromregex`
`numpy.fromstring`
`numpy.full`	`cupynumeric.full`	✓	✓
`numpy.generic`
`numpy.genfromtxt`
`numpy.get_include`
`numpy.get_printoptions`
`numpy.getbufsize`
`numpy.geterr`
`numpy.geterrcall`
`numpy.geterrobj`
`numpy.half`
`numpy.hamming`	`cupynumeric.hamming`	✓	✓
`numpy.hanning`	`cupynumeric.hanning`	✓	✓
`numpy.identity`	`cupynumeric.identity`	✓	✓
`numpy.iinfo`
`numpy.indices`	`cupynumeric.indices`	✓	✓
`numpy.inexact`
`numpy.info`
`numpy.int16`
`numpy.int32`
`numpy.int64`
`numpy.int8`
`numpy.int_`
`numpy.intc`
`numpy.integer`
`numpy.intp`
`numpy.isfortran`
`numpy.isscalar`	`cupynumeric.isscalar`	✓	✓
`numpy.issctype`
`numpy.issubclass_`
`numpy.issubdtype`
`numpy.issubsctype`
`numpy.kaiser`	`cupynumeric.kaiser`	✓	✓
`numpy.load`	`cupynumeric.load`	❌	❌
`numpy.loadtxt`
`numpy.longcomplex`
`numpy.longdouble`
`numpy.longfloat`
`numpy.longlong`
`numpy.lookfor`
`numpy.mask_indices`	`cupynumeric.mask_indices`	✓	✓
`numpy.mat`
`numpy.matrix`
`numpy.maximum_sctype`
`numpy.memmap`
`numpy.mintypecode`
`numpy.ndarray()`
`numpy.ndenumerate`
`numpy.ndindex`
`numpy.nditer`
`numpy.nested_iters`
`numpy.number`
`numpy.obj2sctype`
`numpy.object_`
`numpy.ones`	`cupynumeric.ones`	✓	✓
`numpy.poly1d`
`numpy.printoptions`
`numpy.promote_types`
`numpy.recarray`
`numpy.record`
`numpy.require`
`numpy.sctype2char`
`numpy.set_printoptions`
`numpy.set_string_function`
`numpy.setbufsize`
`numpy.seterr`
`numpy.seterrcall`
`numpy.seterrobj`
`numpy.short`
`numpy.show_config`
`numpy.show_runtime`
`numpy.signedinteger`
`numpy.single`
`numpy.singlecomplex`
`numpy.source`
`numpy.str_`
`numpy.string_`
`numpy.timedelta64`
`numpy.tri`	`cupynumeric.tri`	✓	✓
`numpy.tril_indices`	`cupynumeric.tril_indices`	✓	✓
`numpy.triu_indices`	`cupynumeric.triu_indices`	✓	✓
`numpy.typename`
`numpy.ubyte`
`numpy.ufunc`
`numpy.uint`
`numpy.uint16`
`numpy.uint32`
`numpy.uint64`
`numpy.uint8`
`numpy.uintc`
`numpy.uintp`
`numpy.ulonglong`
`numpy.unicode_`
`numpy.unsignedinteger`
`numpy.ushort`
`numpy.vectorize`
`numpy.void`
`numpy.who`
`numpy.zeros`	`cupynumeric.zeros`	✓	✓

Summary

Number of NumPy functions: 164

Number of functions covered by cupynumeric: 24

Ufuncs#

NumPy	cupynumeric	single-GPU/CPU	multi-GPU/CPU
`numpy.absolute`	`cupynumeric.absolute`	✓	✓
`numpy.add`	`cupynumeric.add`	✓	✓
`numpy.arccos`	`cupynumeric.arccos`	✓	✓
`numpy.arccosh`	`cupynumeric.arccosh`	✓	✓
`numpy.arcsin`	`cupynumeric.arcsin`	✓	✓
`numpy.arcsinh`	`cupynumeric.arcsinh`	✓	✓
`numpy.arctan`	`cupynumeric.arctan`	✓	✓
`numpy.arctan2`	`cupynumeric.arctan2`	✓	✓
`numpy.arctanh`	`cupynumeric.arctanh`	✓	✓
`numpy.bitwise_and`	`cupynumeric.bitwise_and`	✓	✓
`numpy.bitwise_or`	`cupynumeric.bitwise_or`	✓	✓
`numpy.bitwise_xor`	`cupynumeric.bitwise_xor`	✓	✓
`numpy.cbrt`	`cupynumeric.cbrt`	✓	✓
`numpy.ceil`	`cupynumeric.ceil`	✓	✓
`numpy.conj`	`cupynumeric.conj`	✓	✓
`numpy.conjugate`	`cupynumeric.conjugate`	✓	✓
`numpy.copysign`	`cupynumeric.copysign`	✓	✓
`numpy.cos`	`cupynumeric.cos`	✓	✓
`numpy.cosh`	`cupynumeric.cosh`	✓	✓
`numpy.deg2rad`	`cupynumeric.deg2rad`	✓	✓
`numpy.degrees`	`cupynumeric.degrees`	✓	✓
`numpy.divide`	`cupynumeric.divide`	✓	✓
`numpy.divmod`
`numpy.equal`	`cupynumeric.equal`	✓	✓
`numpy.exp`	`cupynumeric.exp`	✓	✓
`numpy.exp2`	`cupynumeric.exp2`	✓	✓
`numpy.expm1`	`cupynumeric.expm1`	✓	✓
`numpy.fabs`	`cupynumeric.fabs`	✓	✓
`numpy.float_power`	`cupynumeric.float_power`	✓	✓
`numpy.floor`	`cupynumeric.floor`	✓	✓
`numpy.floor_divide`	`cupynumeric.floor_divide`	✓	✓
`numpy.fmax`	`cupynumeric.fmax`	✓	✓
`numpy.fmin`	`cupynumeric.fmin`	✓	✓
`numpy.fmod`	`cupynumeric.fmod`	✓	✓
`numpy.frexp`	`cupynumeric.frexp`	✓	✓
`numpy.gcd`	`cupynumeric.gcd`	✓	✓
`numpy.greater`	`cupynumeric.greater`	✓	✓
`numpy.greater_equal`	`cupynumeric.greater_equal`	✓	✓
`numpy.heaviside`
`numpy.hypot`	`cupynumeric.hypot`	✓	✓
`numpy.invert`	`cupynumeric.invert`	✓	✓
`numpy.isfinite`	`cupynumeric.isfinite`	✓	✓
`numpy.isinf`	`cupynumeric.isinf`	✓	✓
`numpy.isnan`	`cupynumeric.isnan`	✓	✓
`numpy.isnat`
`numpy.lcm`	`cupynumeric.lcm`	✓	✓
`numpy.ldexp`	`cupynumeric.ldexp`	✓	✓
`numpy.left_shift`	`cupynumeric.left_shift`	✓	✓
`numpy.less`	`cupynumeric.less`	✓	✓
`numpy.less_equal`	`cupynumeric.less_equal`	✓	✓
`numpy.log`	`cupynumeric.log`	✓	✓
`numpy.log10`	`cupynumeric.log10`	✓	✓
`numpy.log1p`	`cupynumeric.log1p`	✓	✓
`numpy.log2`	`cupynumeric.log2`	✓	✓
`numpy.logaddexp`	`cupynumeric.logaddexp`	✓	✓
`numpy.logaddexp2`	`cupynumeric.logaddexp2`	✓	✓
`numpy.logical_and`	`cupynumeric.logical_and`	✓	✓
`numpy.logical_not`	`cupynumeric.logical_not`	✓	✓
`numpy.logical_or`	`cupynumeric.logical_or`	✓	✓
`numpy.logical_xor`	`cupynumeric.logical_xor`	✓	✓
`numpy.matmul`	`cupynumeric.matmul`	✓	✓
`numpy.maximum`	`cupynumeric.maximum`	✓	✓
`numpy.minimum`	`cupynumeric.minimum`	✓	✓
`numpy.mod`	`cupynumeric.mod`	✓	✓
`numpy.modf`	`cupynumeric.modf`	✓	✓
`numpy.multiply`	`cupynumeric.multiply`	✓	✓
`numpy.negative`	`cupynumeric.negative`	✓	✓
`numpy.nextafter`	`cupynumeric.nextafter`	✓	✓
`numpy.not_equal`	`cupynumeric.not_equal`	✓	✓
`numpy.positive`	`cupynumeric.positive`	✓	✓
`numpy.power`	`cupynumeric.power`	✓	✓
`numpy.rad2deg`	`cupynumeric.rad2deg`	✓	✓
`numpy.radians`	`cupynumeric.radians`	✓	✓
`numpy.reciprocal`	`cupynumeric.reciprocal`	✓	✓
`numpy.remainder`	`cupynumeric.remainder`	✓	✓
`numpy.right_shift`	`cupynumeric.right_shift`	✓	✓
`numpy.rint`	`cupynumeric.rint`	✓	✓
`numpy.sign`	`cupynumeric.sign`	✓	✓
`numpy.signbit`	`cupynumeric.signbit`	✓	✓
`numpy.sin`	`cupynumeric.sin`	✓	✓
`numpy.sinh`	`cupynumeric.sinh`	✓	✓
`numpy.spacing`
`numpy.sqrt`	`cupynumeric.sqrt`	✓	✓
`numpy.square`	`cupynumeric.square`	✓	✓
`numpy.subtract`	`cupynumeric.subtract`	✓	✓
`numpy.tan`	`cupynumeric.tan`	✓	✓
`numpy.tanh`	`cupynumeric.tanh`	✓	✓
`numpy.true_divide`	`cupynumeric.true_divide`	✓	✓
`numpy.trunc`	`cupynumeric.trunc`	✓	✓

Summary

Number of NumPy functions: 89

Number of functions covered by cupynumeric: 85

Multi-Dimensional Array#

NumPy	cupynumeric	single-GPU/CPU	multi-GPU/CPU
`numpy.ndarray.all()`	`cupynumeric.ndarray.all()`	✓	✓
`numpy.ndarray.any()`	`cupynumeric.ndarray.any()`	✓	✓
`numpy.ndarray.argmax()`	`cupynumeric.ndarray.argmax()`	✓	✓
`numpy.ndarray.argmin()`	`cupynumeric.ndarray.argmin()`	✓	✓
`numpy.ndarray.argpartition()`	`cupynumeric.ndarray.argpartition()`	✓	✓
`numpy.ndarray.argsort()`	`cupynumeric.ndarray.argsort()`	✓	✓
`numpy.ndarray.astype()`	`cupynumeric.ndarray.astype()`	✓	✓
`numpy.ndarray.byteswap()`
`numpy.ndarray.choose()`	`cupynumeric.ndarray.choose()`	✓	✓
`numpy.ndarray.clip()`	`cupynumeric.ndarray.clip()`	✓	✓
`numpy.ndarray.compress()`	`cupynumeric.ndarray.compress()`	✓	✓
`numpy.ndarray.conj()`	`cupynumeric.ndarray.conj()`	✓	✓
`numpy.ndarray.conjugate()`	`cupynumeric.ndarray.conjugate()`	✓	✓
`numpy.ndarray.copy()`	`cupynumeric.ndarray.copy()`	✓	✓
`numpy.ndarray.cumprod()`	`cupynumeric.ndarray.cumprod()`	❌	❌
`numpy.ndarray.cumsum()`	`cupynumeric.ndarray.cumsum()`	❌	❌
`numpy.ndarray.diagonal()`	`cupynumeric.ndarray.diagonal()`	✓	✓
`numpy.ndarray.dot()`	`cupynumeric.ndarray.dot()`	✓	✓
`numpy.ndarray.dump()`	`cupynumeric.ndarray.dump()`	❌	❌
`numpy.ndarray.dumps()`	`cupynumeric.ndarray.dumps()`	✓	✓
`numpy.ndarray.fill()`	`cupynumeric.ndarray.fill()`	✓	✓
`numpy.ndarray.flatten()`	`cupynumeric.ndarray.flatten()`	✓	✓
`numpy.ndarray.getfield()`	`cupynumeric.ndarray.getfield()`	❌	❌
`numpy.ndarray.item()`	`cupynumeric.ndarray.item()`	✓	✓
`numpy.ndarray.itemset()`	`cupynumeric.ndarray.itemset()`	✓	✓
`numpy.ndarray.mean()`	`cupynumeric.ndarray.mean()`	✓	✓
`numpy.ndarray.newbyteorder()`
`numpy.ndarray.nonzero()`	`cupynumeric.ndarray.nonzero()`	✓	✓
`numpy.ndarray.partition()`	`cupynumeric.ndarray.partition()`	✓	✓
`numpy.ndarray.prod()`	`cupynumeric.ndarray.prod()`	✓	✓
`numpy.ndarray.ptp()`
`numpy.ndarray.put()`	`cupynumeric.ndarray.put()`	✓	✓
`numpy.ndarray.ravel()`	`cupynumeric.ndarray.ravel()`	✓	✓
`numpy.ndarray.repeat()`
`numpy.ndarray.reshape()`	`cupynumeric.ndarray.reshape()`	✓	✓
`numpy.ndarray.resize()`
`numpy.ndarray.searchsorted()`	`cupynumeric.ndarray.searchsorted()`	✓	✓
`numpy.ndarray.setfield()`	`cupynumeric.ndarray.setfield()`	❌	❌
`numpy.ndarray.setflags()`	`cupynumeric.ndarray.setflags()`	✓	✓
`numpy.ndarray.sort()`	`cupynumeric.ndarray.sort()`	✓	✓
`numpy.ndarray.squeeze()`	`cupynumeric.ndarray.squeeze()`	✓	✓
`numpy.ndarray.std()`
`numpy.ndarray.sum()`	`cupynumeric.ndarray.sum()`	✓	✓
`numpy.ndarray.swapaxes()`	`cupynumeric.ndarray.swapaxes()`	✓	✓
`numpy.ndarray.take()`	`cupynumeric.ndarray.take()`	✓	✓
`numpy.ndarray.tobytes()`	`cupynumeric.ndarray.tobytes()`	✓	✓
`numpy.ndarray.tofile()`	`cupynumeric.ndarray.tofile()`	❌	❌
`numpy.ndarray.tolist()`	`cupynumeric.ndarray.tolist()`	✓	✓
`numpy.ndarray.tostring()`	`cupynumeric.ndarray.tostring()`	✓	✓
`numpy.ndarray.trace()`	`cupynumeric.ndarray.trace()`	✓	✓
`numpy.ndarray.transpose()`	`cupynumeric.ndarray.transpose()`	✓	✓
`numpy.ndarray.var()`	`cupynumeric.ndarray.var()`	✓	✓
`numpy.ndarray.view()`	`cupynumeric.ndarray.view()`	✓	✓

Summary

Number of NumPy functions: 53

Number of functions covered by cupynumeric: 47

Linear Algebra#

NumPy	cupynumeric	single-GPU/CPU	multi-GPU/CPU
`numpy.linalg.LinAlgError`

Summary

Number of NumPy functions: 1

Number of functions covered by cupynumeric: 0

Discrete Fourier Transform#

NumPy	cupynumeric	single-GPU/CPU	multi-GPU/CPU
`numpy.fft.fftfreq`
`numpy.fft.rfftfreq`

Summary

Number of NumPy functions: 2

Number of functions covered by cupynumeric: 0

Random Sampling#

NumPy	cupynumeric	single-GPU/CPU	multi-GPU/CPU
`numpy.random.BitGenerator`
`numpy.random.Generator`
`numpy.random.MT19937`
`numpy.random.PCG64`
`numpy.random.PCG64DXSM`
`numpy.random.Philox`
`numpy.random.RandomState`
`numpy.random.SFC64`
`numpy.random.SeedSequence`
`numpy.random.beta`	`cupynumeric.random.beta`	✓	✓
`numpy.random.binomial`	`cupynumeric.random.binomial`	✓	✓
`numpy.random.bytes`	`cupynumeric.random.bytes`	✓	✓
`numpy.random.chisquare`	`cupynumeric.random.chisquare`	✓	✓
`numpy.random.choice`
`numpy.random.default_rng`	`cupynumeric.random.default_rng`	✓	✓
`numpy.random.dirichlet`
`numpy.random.exponential`	`cupynumeric.random.exponential`	✓	✓
`numpy.random.f`	`cupynumeric.random.f`	✓	✓
`numpy.random.gamma`	`cupynumeric.random.gamma`	✓	✓
`numpy.random.geometric`	`cupynumeric.random.geometric`	✓	✓
`numpy.random.get_bit_generator`
`numpy.random.get_state`
`numpy.random.gumbel`	`cupynumeric.random.gumbel`	✓	✓
`numpy.random.hypergeometric`	`cupynumeric.random.hypergeometric`	✓	✓
`numpy.random.laplace`	`cupynumeric.random.laplace`	✓	✓
`numpy.random.logistic`	`cupynumeric.random.logistic`	✓	✓
`numpy.random.lognormal`	`cupynumeric.random.lognormal`	✓	✓
`numpy.random.logseries`	`cupynumeric.random.logseries`	✓	✓
`numpy.random.multinomial`
`numpy.random.multivariate_normal`
`numpy.random.negative_binomial`	`cupynumeric.random.negative_binomial`	✓	✓
`numpy.random.noncentral_chisquare`	`cupynumeric.random.noncentral_chisquare`	✓	✓
`numpy.random.noncentral_f`	`cupynumeric.random.noncentral_f`	✓	✓
`numpy.random.normal`	`cupynumeric.random.normal`	✓	✓
`numpy.random.pareto`	`cupynumeric.random.pareto`	✓	✓
`numpy.random.permutation`
`numpy.random.poisson`	`cupynumeric.random.poisson`	✓	✓
`numpy.random.power`	`cupynumeric.random.power`	✓	✓
`numpy.random.rand`	`cupynumeric.random.rand`	✓	✓
`numpy.random.randint`	`cupynumeric.random.randint`	✓	✓
`numpy.random.randn`	`cupynumeric.random.randn`	✓	✓
`numpy.random.random`	`cupynumeric.random.random`	✓	✓
`numpy.random.random_integers`	`cupynumeric.random.random_integers`	✓	✓
`numpy.random.random_sample`	`cupynumeric.random.random_sample`	✓	✓
`numpy.random.ranf`	`cupynumeric.random.ranf`	✓	✓
`numpy.random.rayleigh`	`cupynumeric.random.rayleigh`	✓	✓
`numpy.random.sample`	`cupynumeric.random.sample`	✓	✓
`numpy.random.seed`	`cupynumeric.random.seed`	✓	✓
`numpy.random.set_bit_generator`
`numpy.random.set_state`
`numpy.random.shuffle`
`numpy.random.standard_cauchy`	`cupynumeric.random.standard_cauchy`	✓	✓
`numpy.random.standard_exponential`	`cupynumeric.random.standard_exponential`	✓	✓
`numpy.random.standard_gamma`	`cupynumeric.random.standard_gamma`	✓	✓
`numpy.random.standard_normal`
`numpy.random.standard_t`	`cupynumeric.random.standard_t`	✓	✓
`numpy.random.triangular`	`cupynumeric.random.triangular`	✓	✓
`numpy.random.uniform`	`cupynumeric.random.uniform`	✓	✓
`numpy.random.vonmises`	`cupynumeric.random.vonmises`	✓	✓
`numpy.random.wald`	`cupynumeric.random.wald`	✓	✓
`numpy.random.weibull`	`cupynumeric.random.weibull`	✓	✓
`numpy.random.zipf`	`cupynumeric.random.zipf`	✓	✓

Summary

Number of NumPy functions: 62

Number of functions covered by cupynumeric: 42