Add aten::histogram and variant

[majing921201]

CPU/MPS only ops

[majing921201]

two cases failed:
SampleInput(input=tensor([-6.7103, 5.7013, 0.7920, 2.8814, -4.1030], device='xpu:0'), args=(tensor([-6.7103, 5.7013, 0.7920], device='xpu:0'),), kwargs={'weight': None, 'density': False}, broadcasts_input=False, name='')

SampleInput(input=6.362441917740258, args=(tensor([ 6.3624, -3.1650, 7.6229, 4.4501], device='xpu:0',
dtype=torch.float64),), kwargs={'weight': None, 'density': False}, broadcasts_input=False, name='')

cases are not reasonable, for out-of-order bin.
1, pytorch doc requires increasing bin

2, CPU implemented kernel with std::upper_bound, which also requires increasing sequence

[majing921201]

pytorch/pytorch#130916