.NET: Scalability of a Common Dictionary

Question

I use Dictionary<>bazillion to store items. Is it possible to assume that as long as the server memory has enough space to accommodate these bazillion elements, I will get about O (1) extraction of elements from it? What should I know about using a shared dictionary as a huge cache when performance is important?

EDIT: Shouldn't I rely on default implementations? What does a good hashing function do?

Answer 1

It all depends on how good the hashing functionality that your "bazillion items" supports - if their hashing function is not excellent (so many conflicts lead to results), your performance will deteriorate as the dictionary grows.

Answer 2

You must measure it and find out. This is someone who knows the exact use of your vocabulary, so you can measure it to make sure it matches your needs.

: , , . , , , . , .., . , ; , , , .

; , bazillion-item, , , , , .

Answer 3

, O (1). g O (1). , , , GetHashCode , , Equals.

, : , , "" -.

Answer 4

Yes, you will have about O (1) no matter how many objects you put in the dictionary. But for the dictionary to be fast, your key objects must provide a sufficient implementation of GetHashCode, because the dictionary uses a hash table inside itself.