Esempio n. 1
0

@cc_nrt.export('zero_scalar', 'f8(i4)')
def zero_scalar(n):
    arr = np.zeros(n)
    return arr[-1]


if has_blas:
    # This one also needs BLAS
    @cc_nrt.export('vector_dot', 'f8(i4)')
    def vector_dot(n):
        a = np.linspace(1, n, n)
        return np.dot(a, a)


# This one needs an environment
@cc_nrt.export('zeros', 'f8[:](i4)')
def zeros(n):
    return np.zeros(n)


#
# Legacy API
#

exportmany(['multf f4(f4,f4)', 'multi i4(i4,i4)'])(mult)
# Needs to link to helperlib to due with complex arguments
# export('multc c16(c16,c16)')(mult)
export('mult f8(f8, f8)')(mult)
Esempio n. 2
0
#!/usr/bin/env python
from numba.pycc import exportmany, export


def mult(a, b):
    return a * b


export('multi i4(i4, i4)')(mult)
exportmany(['multf f4(f4, f4)', 'mult f8(f8, f8)'])(mult)
Esempio n. 3
0
# This one clashes with libc random() unless pycc is careful with naming.
@cc_helperlib.export('random', 'f8(i4)')
def random_impl(seed):
    if seed != -1:
        np.random.seed(seed)
    return np.random.random()

# These ones need NRT
cc_nrt = CC('pycc_test_nrt')

@cc_nrt.export('zero_scalar', 'f8(i4)')
def zero_scalar(n):
    arr = np.zeros(n)
    return arr[-1]

# This one needs an environment
@cc_nrt.export('zeros', 'f8[:](i4)')
def zeros(n):
    return np.zeros(n)


#
# Legacy API
#

exportmany(['multf f4(f4,f4)', 'multi i4(i4,i4)'])(mult)
# Needs to link to helperlib to due with complex arguments
# export('multc c16(c16,c16)')(mult)
export('mult f8(f8, f8)')(mult)
@cc_nrt.export("zeros", "f8[:](i4)")
def zeros(n):
    return np.zeros(n)


# requires list dtor, #issue3535
@cc_nrt.export("np_argsort", "intp[:](float64[:])")
def np_argsort(arr):
    return np.argsort(arr)


#
# Legacy API
#

exportmany(["multf f4(f4,f4)", "multi i4(i4,i4)"])(mult)
# Needs to link to helperlib to due with complex arguments
# export('multc c16(c16,c16)')(mult)
export("mult f8(f8, f8)")(mult)


@cc_nrt.export("dict_usecase", "intp[:](intp[:])")
def dict_usecase(arr):
    d = typed.Dict()
    for i in range(arr.size):
        d[i] = arr[i]
    out = np.zeros_like(arr)
    for k, v in d.items():
        out[k] = k * v
    return out