Strange behavior while using np.dot with mixed precision arrays in NumPy 1.23
I've tried everything I can think of but I'm relatively new to this, so bear with me... I'm running into an scenario with matrix multiplication using `np.dot` in NumPy 1.23 when my input arrays have mixed precision types. I expected the operation to handle type promotion automatically, but I am getting unexpected results, particularly with integer and float arrays. Here's a minimal example that demonstrates the question: ```python import numpy as np a = np.array([[1, 2], [3, 4]], dtype=np.int32) b = np.array([[1.0, 2.0], [3.0, 4.0]], dtype=np.float64) result = np.dot(a, b) print(result) ``` In this example, I expect the output to be: ``` [ 7. 10.] [15. 22.] ``` However, what I receive is: ``` [ 7. 10.] [15. 22.] ``` To troubleshoot, I explicitly checked the shapes and types of both arrays, and they seem correct. I also tried replacing `np.dot` with `@`, but the output remains unchanged. I verified that the scenario is not related to array shapes by testing simpler cases, and all produce the expected results when using homogeneous types. I checked the [NumPy documentation](https://numpy.org/doc/stable/reference/generated/numpy.dot.html) regarding type promotion, but need to find any mention of issues with mixed precision that would lead to this behavior. Is there something Iām overlooking, or is this a known limitation or behavior in this version? Any insights would be greatly appreciated! I'm working on a web app that needs to handle this. Is this even possible?