SQLite developer Richard Hipp posted recently that the database has become 50% faster than it was 16 months ago. According to the SQLite blog, the database engine, tested with the "'speedtest1 --size 5' workload on Ubuntu 10.13 and gcc 4.8.1 with -Os", now performs 50% more tasks in the same amount of CPU cycles.
What's really interesting is that no one improvement significantly changed SQLite's speed. A combination of plenty of low-end improvements made it work more efficiently, with single changes affecting speed only fractions of a percent.
The 50% faster number above is not about better query plans. This is 50% faster at the low-level grunt work of moving bits on and off disk and search b-trees. We have achieved this by incorporating hundreds of micro-optimizations. Each micro-optimization might improve the performance by as little as 0.05%. If we get one that improves performance by 0.25%, that is considered a huge win. Each of these optimizations is unmeasurable on a real-world system (we have to use cachegrind to get repeatable run-times) but if you do enough of them, they add up.
One-tenth of the improvements to speed have come since the last update, and all tests pass on the new release, so, as Hipp says, "we are confident that we didn't break too much."