Bug description
The call math.exp2DType.float64, 1 works with single-precision floating-point precision.
Steps to reproduce
>>> print(math.exp2[DType.float64, 1](-1022))
1.1754943508222875e-38
>>> print(math.exp2[DType.float32, 1](-126))
1.1754943508222875e-38
>>> print(2.0 ** -1022)
2.2250738585072014e-308