I'm running my cardamon tests on different machines, and there seems to be a mismatch between the [cpu] section and the rest of the file.
I obviously have cardamon.toml checked into git along with my setup and testing code, but I have to locally change the curves for different machines, leading to eternal git conflicts.
An idea:
Named cpu's: Allow multiple cpu sections, and the runtime will pick the appropriate one. Either the runtime detects the right cpu, or as an (optional) parameter.