Datatype sizes

Integers

• int8
• int16
• int32 (int for 32bits)
• int64 (int for 64bits)
• uint8
• uint16
• uint32
• uint64

np.array([2**32], dtype=np.int32).dtype
# shows deprecation warning because of out-of-bound integer
# conversion to smaller data size

We can check the size of numpy integer datatype using method np.iinfo.

np.iinfo(np.int32)
# iinfo(min=-2147483648, max=2147483647, dtype=int32)

Floats

To check the size of float integers, we can use method np.finfo.

np.finfo(np.float64)
# finfo(resolution=1e-15, min=-1.7976931348623157e+308, max=1.7976931348623157e+308, dtype=float64)

Available floats,

• float16
• float32
• float64
• float96 platform dependent (or np.longdouble)
• float128 platform dependent (or np.longdouble)

np.finfo(np.float32).eps
# 1.1920929e-07

Note

eps is machine epsilon which defines round off error for float numbers.

Complex numbers

• complex64 two $32$ bits floats
• complex128 two $64$ bits floats
• complex192 two $96$ bits floats, platform depedent
• complex256 two $128$ bits floats, platform depedent

References

1. https://lectures.scientific-python.org/intro/numpy/elaborate_arrays.html#different-data-type-sizes