Routines

The following pages describe NumPy-compatible routines. These functions cover a subset of NumPy routines.

• Array Creation Routines
  • Basic creation routines
  • Creation from other data
  • Numerical ranges
  • Matrix creation
• Array Manipulation Routines
  • Basic operations
  • Changing array shape
  • Transpose-like operations
  • Changing number of dimensions
  • Changing kind of array
  • Joining arrays
  • Splitting arrays
  • Tiling arrays
  • Adding and removing elements
  • Rearranging elements
• Binary Operations
  • Elementwise bit operations
  • Bit packing
  • Output formatting
• Data Type Routines
  • cupy.can_cast
  • cupy.result_type
  • cupy.common_type
  • Creating data types
  • Data type information
  • Data type testing
  • Miscellaneous
• FFT Functions
  • Standard FFTs
  • Real FFTs
  • Hermitian FFTs
  • Helper routines
  • Normalization
  • Code compatibility features
  • Multi-GPU FFT
• Functional programming
  • cupy.piecewise
• Indexing Routines
  • cupy.c_
  • cupy.r_
  • cupy.nonzero
  • cupy.where
  • cupy.indices
  • cupy.ix_
  • cupy.ravel_multi_index
  • cupy.unravel_index
  • cupy.take
  • cupy.take_along_axis
  • cupy.choose
  • cupy.compress
  • cupy.diag
  • cupy.diag_indices
  • cupy.diag_indices_from
  • cupy.diagonal
  • cupy.extract
  • cupy.select
  • cupy.lib.stride_tricks.as_strided
  • cupy.place
  • cupy.put
  • cupy.putmask
  • cupy.fill_diagonal
  • cupy.flatiter
• Input and Output
  • NumPy binary files (NPY, NPZ)
  • String formatting
  • Base-n representations
• Linear Algebra
  • Matrix and vector products
  • Decompositions
  • Matrix eigenvalues
  • Norms etc.
  • Solving linear equations
  • Special Matrices
• Logic Functions
  • Truth value testing
  • Infinities and NaNs
  • Array type testing
  • Logic operations
  • Comparison
• Mathematical Functions
  • Trigonometric functions
  • Hyperbolic functions
  • Rounding
  • Sums, products, differences
  • Exponents and logarithms
  • Other special functions
  • Floating point routines
  • Arithmetic operations
  • Handling complex numbers
  • Miscellaneous
• Padding
  • cupy.pad
• Polynomials
  • Polynomial Package
  • Poly1d
• Random Sampling (cupy.random)
  • Simple random data
  • Permutations
  • Distributions
  • Random generator
• Sorting, Searching, and Counting
  • Sorting
  • Searching
  • Counting
• Statistical Functions
  • Order statistics
  • Means and variances
  • Histograms
  • Correlations
• CuPy-specific Functions
  • cupyx.rsqrt
  • cupyx.scatter_add
  • cupyx.scatter_max
  • cupyx.scatter_min

CUB/cuTENSOR backend for reduction routines

Some CuPy reduction routines, including sum(), amin(), amax(), argmin(), argmax(), and other functions built on top of them, can be accelerated by switching to the CUB or cuTENSOR backend. These backends can be enabled by setting CUPY_ACCELERATORS environement variable as documented here. Note that while in general the accelerated reductions are faster, there could be exceptions depending on the data layout. We recommend users to perform some benchmarks to determine whether CUB/cuTENSOR offers better performance or not.