11 ms·

Open source Database Breakthrough: 10-80 times faster

6 points by messel 17y ago

russell 17y ago
It gets the speedup by loading sets of column values into CPU cache and performing relational operators on the cache values. Main memory serves a a pipeline to the cache.
- messel 17y ago
  I like it when HN database fluent folks shrink down a post into a one liner. Nice russell.
bayareaguy 17y ago
It would be interesting to see how this compares to MonetDB[1] (another open source db that does operations column-wise), given that it too is often 10x faster (and for some queries nearly 80x faster) than PostgreSQL or MySQL[2]. 1- http://monetdb.cwi.nl/projects/monetdb http://monetdb.cwi.nl/projects/monetdb 2- http://monetdb.cwi.nl/projects/monetdb/SQL/Benchmark/TPCH/ http://monetdb.cwi.nl/projects/monetdb/SQL/Benchmark/TPCH/