New open-source blqsort algorithm outperforms C++ std::sort on benchmarks
A developer released blqsort, a quicksort implementation available on GitHub under the MIT license, with both C++ and C versions including multithreaded variants. Benchmarks on Apple M1 and Intel i5-9300H systems show blqsort sorting 50 million random doubles, partially sorted data, and structs faster than both std::sort and the ipnsort ClickHouse C++ port.
GoKawiil's interpretation of the reporting above, not reported fact.
Faster general-purpose sorting could speed up data-heavy applications that rely on standard library sort functions without requiring specialized SIMD code, according to the developer's claims. The results suggest techniques like branchless partitioning and sorting networks for small partitions can meaningfully beat widely used implementations, though independent verification beyond the author's own benchmarks is not mentioned.
- blqsort reportedly beats std::sort and ipnsort across random, partially sorted, and struct-based data tests.
- The algorithm is open source under MIT license, with single-header C and C++ versions, including multithreaded options.
- Multithreaded blqsort on Apple M1 is claimed to run roughly three times faster than the single-threaded version.
Source: tiki.li — Christof Kaser, 2026-11-10
Published there as: “A C++ Quicksort Faster Than std:sort”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.