The bug in ldm shouldn't have happened before the latest version was published. Since this is an NP complete problem though, it's not guaranteed that one solution should be better than another—all of which are going to be algoritm dependent. There should be known solutions for each algo out there though, so it'd be best to find a set for testing.
There are a few sets here which can be somewhat helpful.
[771,121,281,854,885,734,486,1003,83,62] for example has an objective of 0, 26, 32 for bruteForce, largestDifference and greedy respectively. May be useful to test the anytime on this too.
The bug in ldm shouldn't have happened before the latest version was published. Since this is an NP complete problem though, it's not guaranteed that one solution should be better than another—all of which are going to be algoritm dependent. There should be known solutions for each algo out there though, so it'd be best to find a set for testing.