ann-benchmarks
RoaringBitmap
Our great sponsors
ann-benchmarks | RoaringBitmap | |
---|---|---|
50 | 24 | |
4,568 | 3,372 | |
- | 1.2% | |
8.1 | 8.5 | |
3 days ago | 7 days ago | |
Python | Java | |
MIT License | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
ann-benchmarks
- ANN Benchmarks
-
Approximate Nearest Neighbors Oh Yeah
https://ann-benchmarks.com/ is a good resource covering those libraries and much more.
-
pgvector vs Pinecone: cost and performance
We utilized the ANN Benchmarks methodology, a standard for benchmarking vector databases. Our tests used the dbpedia dataset of 1,000,000 OpenAI embeddings (1536 dimensions) and inner product distance metric for both Pinecone and pgvector.
-
Vector database is not a separate database category
Data warehouses are columnar stores. They are very different from row-oriented databases - like Postgres, MySQL. Operations on columns - e.g., aggregations (mean of a column) are very efficient.
Most vector databases use one of a few different vector indexing libraries - FAISS, hnswlib, and scann (google only) are popular. The newer vector dbs, like weaviate, have introduced their own indexes, but i haven't seen any performance difference -
Reference: https://ann-benchmarks.com/
-
How We Made PostgreSQL a Better Vector Database
(Blog author here). Thanks for the question. In this case the index for both DiskANN and pgvector HNSW is small enough to fit in memory on the machine (8GB RAM), so there's no need to touch the SSD. We plan to test on a config where the index size is larger than memory (we couldn't this time due to limitations in ANN benchmarks [0], the tool we use).
To your question about RAM usage, we provide a graph of index size. When enabling PQ, our new index is 10x smaller than pgvector HNSW. We don't have numbers for HNSWPQ in FAISS yet.
- Do we think about vector dbs wrong?
-
Vector Search with OpenAI Embeddings: Lucene Is All You Need
In terms of "All You Need" for Vector Search, ANN Benchmarks (https://ann-benchmarks.com/) is a good site to review when deciding what you need. As with anything complex, there often isn't a universal solution.
txtai (https://github.com/neuml/txtai) can build indexes with Faiss, Hnswlib and Annoy. All 3 libraries have been around at least 4 years and are mature. txtai also supports storing metadata in SQLite, DuckDB and the next release will support any JSON-capable database supported by SQLAlchemy (Postgres, MariaDB/MySQL, etc).
-
Vector databases: analyzing the trade-offs
pg_vector doesn't perform well compared to other methods, at least according to ANN-Benchmarks (https://ann-benchmarks.com/).
txtai is more than just a vector database. It also has a built-in graph component for topic modeling that utilizes the vector index to autogenerate relationships. It can store metadata in SQLite/DuckDB with support for other databases coming. It has support for running LLM prompts right with the data, similar to a stored procedure, through workflows. And it has built-in support for vectorizing data into vectors.
For vector databases that simply store vectors, I agree that it's nothing more than just a different index type.
-
Vector Dataset benchmark with 1536/768 dim data
The reason https://ann-benchmarks.com is so good, is that we can see a plot of recall vs latency. I can see you have some latency numbers in the leaderboard at the bottom, but it's very difficult to make a decision.
As a practitioner that works with vector databases every day, just latency is meaningless to me, because I need to know if it's fast AND accurate, and what the tradeoff is! You can't have it both ways. So it would be helpful if you showed plots showing this tradeoff, similar to ann-benchmarks.
-
Numbers every LLM Developer should know
There are very efficient algorithms for doing this, but of course it may still be expensive if your dataset is very large. See https://ann-benchmarks.com/ for some of the algorithms
RoaringBitmap
-
Iterating over Bit Sets Quickly
I was recently reading about Roaring https://roaringbitmap.org/ which is a highly optimized compressed bitset implementation. I reccomend reading about it if you are interested in this sort of thing. The talk at https://roaringbitmap.org/talks/ is especially good.
- Roaring Bitmaps
- Roaring bitmaps are compressed bitmaps, can be 100x faster
-
What feature would you like to remove in C++26?
However, I would love compressed (not just packed) bitsets too, which is something different to me. I would make it another class with a similar interface, based on something like roaring. It doesn't need to be in the standard, but it would be nice if the API was a such that one could easily swap implementations.
-
Jaccard Index
As an aside if you find yourself having to compute them on the fly, know that the Roaring Bitmaps libraries is the way to go [1]. The bitmaps are compressed, and can be streamed directly into SIMD computations (batching XORs and popcnts 256 bits wide!). The Jaccard index is just intersection_len / union_len [2] away
[1] https://roaringbitmap.org/
[2] https://roaringbitmap.readthedocs.io/en/latest/#roaringbitma...
-
Looking for fast, space-efficient key-lookup
Use a two stage approach, with a bloom/cuckoo filter stored as a https://roaringbitmap.org/ in memory. Then a secondary key/value store on disk (bolt or anything else).
-
BitSet Vs BigInteger
As an aside, if you're dealing with large bit sets, you might also want to evaluate Roaring Bitmaps.
-
Negative Incentives in Academic Research
Sidetracking a bit the conversation. What a coincidence that the author (Lemire) is also represented on Today's #1 "Ask HN: What are some cool but obscure data structures you know about?" as he is the main contributor of RoaringBitmap https://github.com/RoaringBitmap/RoaringBitmap and one of the main authors of the data structure.
- Ask HN: What are some 'cool' but obscure data structures you know about?
- Roaring bitmaps: A better compressed bitset
What are some alternatives?
pgvector - Open-source vector similarity search for Postgres
HyperMinHash-java - Union, intersection, and set cardinality in loglog space
faiss - A library for efficient similarity search and clustering of dense vectors.
lucene - Apache Lucene open-source search software
Milvus - A cloud-native vector database, storage for next generation AI applications
CQEngine - Ultra-fast SQL-like queries on Java collections
tlsh
Primes - Prime Number Projects in C#/C++/Python
vald - Vald. A Highly Scalable Distributed Vector Search Engine
Feign - Feign makes writing java http clients easier
pgANN - Fast Approximate Nearest Neighbor (ANN) searches with a PostgreSQL database.
maven-compiler-plugin - Apache Maven Compiler Plugin