Revisit if halving of `--max-memory` value is appropriate

The code comment suggests that the halving is done to account for needing to hold both inputs and outputs of a section in memory: https://github.com/DiamondLightSource/httomo/blob/5dbdde134775ff29101c70e6e80c927a7a28381a/httomo/cli.py#L425-L426

However, the function calculating the backing to the dataset store is accounting for the size of both the input and output chunks. Meaning, is there a need to halve the `--max-memory` value if both inputs and outputs are already being accounted for?