'mlx.core.matmul/tensordot' do not support integer type matrix operations

[Redempt1onzzZZ]

Describe the bug
I found that mlx.core.matmul/tensordot do not support integer type matrix operations, but numpy does, and it is not stated in the mlx documentation that these two APIs do not support integer type operations

To Reproduce

Include code snippet

import mlx.core as mx
import numpy as np

a1 = np.array([[1, 2], [3, 4]])
b1 = np.array([[5, 6], [7, 8]])
npresult = np.tensordot(a1, b1)
print('numpy', npresult)

a = mx.array([[1, 2], [3, 4]])
b = mx.array([[5, 6], [7, 8]])
mlxresult = mx.tensordot(a, b)
print('mlx', mlxresult)

import mlx.core as mx
import numpy as np

a1 = np.array([[1, 2], [3, 4]])
b1 = np.array([[5, 6], [7, 8]])
npresult = np.matmul(a1, b1)
print('numpy', npresult)

a = mx.array([[1, 2], [3, 4]])
b = mx.array([[5, 6], [7, 8]])
mlxresult = mx.matmul(a, b)
print('mlx', mlxresult)

Desktop (please complete the following information):

• OS Version: MacOS 14.2.1
• Version 0.10.0
• numpy 1.26.2

Additional context
According to my inspection, only these two APIs were found to have this issue. I am not sure if there are other APIs that also have this issue. If it is the design of mlx itself, I think it can be explained in the documentation.

[awni]

This is not a huge priority for us since it's a decent amount of work for something which is not super useful. We can update the docs for now until someone has bandwidth to add it or makes a contribution.

[ParamThakkar123]

Would type_promotion help here in anyways to solve this issue ?

[awni]

No, we should just implement the integer gemms.

[zcbenz]

I'm closing this as won't fix since we didn't see any other request for integer matmul, and we will likely deny addition of such kernels.