time/1 with number of inferences

[UWN]

Currently time/1 only shows

% CPU time: 13.900 seconds

Ideally, also the number of inferences would be ideal. Often time complexity then becomes much easier to estimate.

[triska]

On the WAM level, is the number of inferences equivalent to the number of call and execute instructions that have been interpreted?

[UWN]

More or less, built-in calls need to be added, probably.

A property, that I would love to be observed exactly is that failure-slices do indeed have less or equal as many inferences as the original program.

[UWN]

There is already call_with_inference_limit/3, so it should not be that difficult.

[triska]

I have filed #1999 for this. However, it does not yet show the inferences when the goal fails. Would it still be useful to add it like this?