Question 1

How does unordered_dense compare to absl::flat_hash_map in terms of speed?

Accepted Answer

Benchmarks indicate unordered_dense is in the same performance ballpark as absl::flat_hash_map, with similar fast insertions and lookups, but it uses a different storage strategy (contiguous vector) that can offer better iteration speed at the cost of slower deletions.

Question 2

Is unordered_dense thread-safe for concurrent access?

Accepted Answer

No, the library does not provide built-in thread safety; like most STL containers, concurrent modifications require external synchronization such as mutexes, which is not mentioned in the README as a feature.

Question 3

How to reduce memory usage with unordered_dense in large datasets?

Accepted Answer

Use the segmented_map or segmented_set variants, which allocate memory in chunks to lower peak usage and provide stable references on insert, though at the cost of slightly slower indexing due to indirection.

Question 4

Can I use unordered_dense in embedded C++ projects with limited RAM?

Accepted Answer

Yes, but caution is needed: the dense storage minimizes overhead, but deletions might be slow on resource-constrained systems. Segmented variants help smooth memory usage, but non-contiguous data could affect cache performance.

Question 5

How to implement a custom hash function for my key type?

Accepted Answer

Define a struct with operator() returning a uint64_t, and optionally add using is_avalanching = void; for high-quality hashes. The README provides examples, such as using wyhash for avalanching effects.

Question 6

What are the trade-offs of robin-hood backward shift deletion in this library?

Accepted Answer

It enables fast lookups and efficient space usage by minimizing collisions, but as the README notes, it makes deletions slower because elements need to be shifted and indexes updated, requiring extra computations.

Question 7

unordered_dense or std::unordered_map for a beginner's project?

Accepted Answer

Stick with std::unordered_map for simplicity: unordered_dense's performance gains come with complexities like iterator instability and custom hashing, which might overwhelm new developers unless performance is critical.

unordered_dense

What is unordered_dense?

Overview

Use Cases

Best For

Related Projects

Found a gem we're missing?

Not Ideal For

Pros & Cons

Pros

Cons

Frequently Asked Questions