How to enable concurrency in queries?

My collection has 100 million documents. There is a counter field with values from 1 to 100 million uniquely assigned to each document and indexed.

If a query needs to be run on this collection to find documents that match certain filter conditions, how do enable currency?

For example, the collection could be divided into 100 parts. And then there would be 100 concurrent threads searching those 100 parts separately.

How can the above be achieved?

I can’t believe such thing exists in any well known DBMS, but let me know if there’s one.

In most cases, all indexes only sit in memory when db is running, so why bother using multiple threads on it?

this is much easier said than done. And i even doubt if this can really make the queries faster.