unordered
hashtable-benchmarks
unordered | hashtable-benchmarks | |
---|---|---|
10 | 8 | |
53 | 29 | |
- | - | |
9.2 | 4.7 | |
7 days ago | 5 months ago | |
C++ | Java | |
Boost Software License 1.0 | 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.
unordered
-
Effortless Performance Improvements in C++: std::unordered_map
We added two new benchmarks to Boost.Unordered, word_count and word_size, and the second one ends up testing a small hash table (114 elements in 64 bit, even fewer in 32 bit because we use a smaller input file there.)
-
Inside boost::unordered_flat_map
Hi, we have seen similar gains with __forceinline in MSVC, looks like this compiler is not particularly aggressive at inlining. Could you please file an issue at Boost.Unordered repo so what we don't forget? Thank you
-
Boost 1.81 will have boost::unordered_flat_map...
You can request a feature be added by opening an issue in https://github.com/boostorg/unordered.
-
boost::unordered map is a new king of data structures
Here are the results of our uint32.cpp synthetic benchmark under VS2022 Release x64:
-
Advancing the state of the art for <code>std::unordered_map</code> implementations
You can run these benchmarks yourself from the Boost develop branch, they are in the Unordered repo. Since Unordered is header-only, there should be need to build Boost but you do need to bootstrap and then run b2 headers to create the symlinks in boost/.
-
New Boost.Unordered containers have BIG improvements!
Make sure you checkout the preview.md for instructions on how to build nightly Boost in a way that's non-intrusive to your system and works with CMake.
hashtable-benchmarks
-
Building a faster hash table for high performance SQL joins
Since the blog post mentioned a PR to replace linear probing with Robin Hood, I just wanted to mention that I found bidirectional linear probing to outperform Robin Hood across the board in my Java integer set benchmarks:
https://github.com/senderista/hashtable-benchmarks/blob/mast...
https://github.com/senderista/hashtable-benchmarks/wiki/64-b...
-
Ask HN: Who wants to be hired? (December 2023)
https://homes.cs.washington.edu/~magda/papers/wang-cidr17.pd...
I'm most interested in developing high-performance database engines in low-level languages, but open to any challenging systems programming project. I've been working in C++ for the last 3 years, but have written nontrivial projects in Rust and Java as well (e.g., https://github.com/senderista/rotated-array-set, https://github.com/senderista/hashtable-benchmarks). I would enjoy using Rust or Zig on a new project, but I consider the project itself to be much more important than the language it's written in. I am not interested in cryptocurrency, adtech, or fintech projects.
-
Factor is faster than Zig
Thanks for the details on your benchmarks. I would like sometime to extend BLP to a more generic setting; as I said I think any trick used with RH would also work with BLP. I just used an integer set because that's all I needed for my use case and it was easy to implement several different approaches for benchmarking. As you note, it favors use cases where the hash function is cheap (or invertible) and elements are cheap to move around.
About your question on load factors: no, the benchmarks are measuring exactly what they claim to be. The hash table constructor divides max data size by load factor to get the table size (https://github.com/senderista/hashtable-benchmarks/blob/mast...), and the benchmark code instantiates each hash table for exactly the measured data set size and load factor (https://github.com/senderista/hashtable-benchmarks/blob/mast...).
I can't explain the peaks around 1M in many of the plots; I didn't investigate them at the time and I don't have time now. It could be a JVM artifact, but I did try to use JMH "best practices", and there's no dynamic memory allocation or GC happening during the benchmark at all. It would be interesting to port these tables to Rust and repeat the measurements with Criterion. For more informative graphs I might try a log-linear approach: divide the intervals between the logarithmically spaced data sizes into a fixed number of subintervals (say 4).
-
Inside boost::unordered_flat_map
I think "bidirectional linear probing" is an underrated approach (and much simpler): https://github.com/senderista/hashtable-benchmarks/blob/master/src/main/java/set/int64/BLPLongHashSet.java
-
A fast & densely stored hashmap and hashset based on robin-hood backward shift deletion
I will probably never get around to porting my bidirectional linear probing integer hash set from Java to C++, but I hope someone can try adapting BLP to general C++ hashmaps and hashsets, because it significantly outperforms Robin Hood in my benchmarks.
-
Ask HN: Who wants to be hired? (March 2022)
https://homes.cs.washington.edu/~magda/papers/wang-cidr17.pd...
I'm most interested in developing high-performance database engines in low-level languages, but open to any challenging systems programming project. I've been working in C++ for the last 2 years, but have written nontrivial projects in Rust and Java as well (e.g., https://github.com/senderista/rotated-array-set, https://github.com/senderista/hashtable-benchmarks). I would enjoy using Rust or Zig on a new project, but I consider the project itself to be much more important than the language it's written in. I am not interested in cryptocurrency, adtech, or fintech projects.
What are some alternatives?
flat_hash_map - A very fast hashtable
unordered_dense - A fast & densely stored hashmap and hashset based on robin-hood backward shift deletion
FetchBoostContent - CMake FetchContent for Boost libraries
myria - Myria is a scalable Analytics-as-a-Service platform based on relational algebra.
js2scheme
Hopscotch map - C++ implementation of a fast hash map and hash set using hopscotch hashing
boostdep - A tool to create Boost module dependency reports
robin-hood-hashing - Fast & memory efficient hashtable based on robin hood hashing for C++11/14/17/20
emhash - Fast and memory efficient c++ flat hash map/set
nafeez.xyz - ⚡ My personal website.