Sc.tl.rank_genes_groups - cell cluster size and cell_cluster vs cell_cluster comparison

Hi guys,
Two questions:
Question1:

Lets say I have an adata object with 3 clusters.
Cluster1 has 100
Cluster2 has 10000
Cluster3 has 100
cells in it.

When I run
sc.tl.rank_genes_groups
on the adata and I get the ranking for the cell specific markers.
The cell cluster size does not have an impact on the overall cell type specificity I assume?

Question2:
Lets take the adata again from above:
This time I want to compare only cluster 1 and cluster 3 for different marker genes?
Would I do a subset of the adata to do that, or is there a way to specify that in the sc.tl.rank_genes_groups function?

Thank you & best wishes,
Simon