Comprehensive C++ Hashmap Benchmarks (2022)(martin.ankerl.com)
martin.ankerl.com
Comprehensive C++ Hashmap Benchmarks (2022)
https://martin.ankerl.com/2022/08/27/hashmap-bench-01/
20 comments
The F14/Abseil maps are included and use cute SIMD tricks, FWIW (discussed by the blogspot author).
Benchmarks are fine but they will only be loosely correlated with the measured performance for any specific use case.
There is still substantial performance to be gained by creating bespoke hashmap designs at every point of use in code. The high dimensionality of the algorithm optimization space makes it improbable that any specific hashmap algorithm implementation will optimally capture the characteristics of a use case or set of use cases. The variance can be relatively high.
It isn't uncommon to find several independent hashmap designs inside performance-engineered code bases. The sensitivity to small details makes it difficult to build excellent hashmap abstractions with broad scope.
There is still substantial performance to be gained by creating bespoke hashmap designs at every point of use in code. The high dimensionality of the algorithm optimization space makes it improbable that any specific hashmap algorithm implementation will optimally capture the characteristics of a use case or set of use cases. The variance can be relatively high.
It isn't uncommon to find several independent hashmap designs inside performance-engineered code bases. The sensitivity to small details makes it difficult to build excellent hashmap abstractions with broad scope.
Definitely. As an extreme but fun example... in one project I had a massive hash map (~700 GB or so) that was concurrently read to/written from by 256 threads. The entries themselves were only 16 bytes and so I could use atomic cmpxchg, but the problem I hit was that even with 1GB huge pages, I was running out of dTLB entries. So I assigned each thread to a subregion of the hash table, then used channels between each pair of threads to handle the reads and writes (and restructured the program a bit to allow this). Since the dTLB budget is per core, this allowed me to get essentially 0 dTLB misses, and ultimately sped up the program by ~2x
The "delegation pattern" for datastructures:
https://timharris.uk/papers/2013-opodis.pdf
https://timharris.uk/papers/2013-opodis.pdf
ah! I thought I was being original :)
It's also the case that performance of a hashmap, or anything, in a small-scale benchmark may not reflect the performance in a large system that does things other than manage maps. There are side effects like how many icache lines are visited during a map operation. These don't hurt microbenchmarks but they can matter in real systems. It may not be completely pointless to microbenchmark a data structure but it can be ultimately misleading.
Totally. It's funny how many people do not actually realize this and get stuck in cargo-cult mindset forever, no matter the years of experience.
The performance of hash tables and hash functions significantly depends on the data distribution, and should be compared on real datasets.
I've covered it in my presentation: https://presentations.clickhouse.com/2017-hash_tables/
I've covered it in my presentation: https://presentations.clickhouse.com/2017-hash_tables/
"Each benchmark was run multiple times, and I’m using the median to get rid of any potential outliers."
This is not how you should do benchmarks. Don't take the median, you don't even need to do any "warming up".
Simply run it long enough and only take the best result of each. This is more reliable and correct.
This is not how you should do benchmarks. Don't take the median, you don't even need to do any "warming up".
Simply run it long enough and only take the best result of each. This is more reliable and correct.
This is not universally applicable, especially if an algo isn't deterministic. For example if you were to time "bogosort of 100 items" you'd see increasingly better times the more runs you performed.
Would be interested to hear how the Qt QHash compares.
https://doc.qt.io/qt-6/qhash.html
https://doc.qt.io/qt-6/qhash.html
Still using linked lists as std::unordered_map. So it won't fly, but keeps ptr stability.
Didn't all the QList<> linked lists become QVector<> is Qt 6?
It's a bit confusing. QVector in Qt 6 is an alias to QList, which has now been changed to work like QVector did.
https://www.youtube.com/watch?v=nSFIFlLpsTQ
https://www.youtube.com/watch?v=nSFIFlLpsTQ
Not really comprehesive. Doesn't include my favorite
https://github.com/greg7mdp/parallel-hashmap which adds thread-safety to performance.
I wanted to mention this because boost::unordered_flat_map and boost::unordered_flat_set are among the fastest open addressing hash containers in C++ land. Internally, they use lots of cool SIMD tricks. If anyone is interested in the details, here's a nice blog post by the developer: https://bannalia.blogspot.com/2022/11/inside-boostunorderedf...