Announcement

Collapse
No announcement yet.

Google Open-Sources C++ B-Tree Containers Library

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Google Open-Sources C++ B-Tree Containers Library

    Phoronix: Google Open-Sources C++ B-Tree Containers Library

    Google has open-sourced C++ B-Tree, their C++ template library for implementing B-tree containers with similar interfaces to the standard STL map/set/multimap/multiset containers...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Useful for BTRFS or did they come up with their own solution?
    All opinions are my own not those of my employer if you know who they are.

    Comment


    • #3
      Originally posted by Ericg View Post
      Useful for BTRFS or did they come up with their own solution?
      Google released a C++ Template Library. Btrfs is kernel code, so C++ is not an opinion.

      Comment


      • #4
        Originally posted by Ericg View Post
        Useful for BTRFS or did they come up with their own solution?
        yes we should all throw out our tried-and-tested algorithms, just because google decides to use the internet as its QA department

        Comment


        • #5
          Comparing this to STL maps is silly. It's like the language shootout where someone brags how their language is only 2x as slow when their code is the best the language can do while C++ has both hands tied behind its back and a leg cut off. There's a reason we don't use STL containers in games. The use cases their design is based on (holding large, expensive to copy objects) doesn't exist in the real world that I can tell. Especially now that we have move semantics. STL unordered_map for instance is one of the worst hash table implementations, because they assume you can't afford to move elements. A sane open-addressing implementation can be two orders of magnitude faster in real work loads for my tests.

          Showing Google's B-tree results against the STL being faster is like saying water is wet. No shit, Sherlock.

          Comment

          Working...
          X