Add task to processor mapping to output

[jschwartz-cray]

IOR 2.x had code to produce output like this on verbose >= VERBOSE_2:

 task 0 on nid00101
 task 1 on nid00101
 task 10 on nid00101
 ...

which was sorted (though it was purely lexicographic hence why 10 shows up before 2) and very useful for debugging data corruption issues since you always want to determine which processor the writer and reader tasks were on.

I plan to submit a PR which adds a TaskToProcessor output section that contains this information, e.g.

TaskToProcessor:
0                   : nid00101
1                   : nid00101
2                   : nid00101
3                   : nid00101
4                   : nid00102
5                   : nid00102
6                   : nid00102
7                   : nid00102
8                   : nid00103
9                   : nid00103
10                  : nid00103
11                  : nid00103
12                  : nid00104
13                  : nid00104
14                  : nid00104
15                  : nid00104