hopperbench-sqlite

Overview ¶

Command hopperbench-sqlite measures hopper's hot paths on SQLite, as hopperbench does on Postgres, and prints its results in hopperbench's format, one JSON object per scenario, which hopperbench -compare reads.

hopperbench-sqlite [-dir /tmp] [-scenarios throughput,workflow,insert,latency] [-jobs 50000] [-clients 4]

Each scenario runs on a database file of its own in a temporary directory under -dir, removed afterwards. -dir must be on a local disk.

Scenarios:

throughput  no-op jobs preloaded, then drained by -clients clients
retry       as throughput, with every job failing its first attempt
workflow    as throughput, in workflows of ten steps (one, then eight after it, then one after those),
            each step recording a small output; -jobs is rounded down to a multiple of ten
insert      InsertMany in batches of 100
latency     commit-to-Work on an idle queue, local wake-up