Runtime attribute settings and documentation for different batch/cohort sizes

[mwalker174]

We should publish working RuntimeAttr inputs for each module at different batch sizes. Note that the WDL defaults are generally for single-batch runs of ~100-200 samples.

Note that batch size is not always a perfect predictor of memory/disk usage. For example, number of variants, allele frequency, and evidence density can also significantly impact resource usage, so sample quality, PCR status, filtering criteria, etc. should also be considered. Documentation for tasks/workflows affected by these factors would also be helpful.