NumPy and cuNumeric behave differently in 1 case for any/all API

[robinwnv]

Case-1

axes=(-1, 0)
input = [[[5, 10], [0, 100]]]
in_num = num.array(input)
fn_num = getattr(num, func)
out_num = fn_num(in_num,  axis=axes)

func='any' or 'all'
In Numpy, it pass
In cuNumeric, it raises ValueError: Invalid promotion on dimension 2 for a 1-D store

[robinwnv]

input = [[[5, 10], [0, 100]]]
np.all(input_a, axis=(-1,0))
array([ True, False])
np.any(input_a, axis=(-1,0))
array([ True, True])
num.all(input, axis=(-1,0))
array([ True, False])
num.any(input, axis=(-1,0))
array([ True, True])

The issue is solved now. So close this issue.