Ok, why would selecting the primary key columns change anything here in your opinion?
Having a different data structure to store your table's data will make a difference if the amount of data (not the amount of records) that needs to be processed is different.
Not knowing your actual data it's hard to guess, but for such few records, it's very likely that the structural overhead of the column store by itself is larger than the possible data compression savings.
Remember that each and every column needs to have it's own dictionary, value vector and maybe inverted indexes.
Concerning response time measurement: the server processing time does not include the time required to transfer data to your application. That may or may not be an issue (e.g. when the query itself is only used to be fed into SQLScript code).
Other than that you didn't provide any more specifics (table definition, data, queries) that would allow a more general answer.