Skip to content

PERF.md

Latest commit

 

History

History
64 lines (38 loc) · 1.54 KB

PERF.md

File metadata and controls

64 lines (38 loc) · 1.54 KB

Performance Testing

Mobibench

  1. Clone the source repository of Mobibench fork for Turso:
git clone git@github.com:penberg/Mobibench.git
  1. Build Mobibench:
cd Mobibench/shell
LIBS="../../target/release/libturso_sqlite3.a -lm" make
mv mobibench mobibench-turso
  1. Run Mobibench:

(easiest way is to cd into target/release)

# with strace, from target/release

strace -f -c ../../Mobibench/shell/mobibench-turso -f 1024 -r 4 -a 0 -y 0 -t 1 -d 0 -n 10000 -j 3 -s 2 -T 3 -D 1


./mobibench -p <benchmark-directory> -n 1000 -d 0 -j 4

Clickbench

We have a modified version of the Clickbench benchmark script that can be run with:

make clickbench

This will build Turso in release mode, create a database, and run the benchmarks with a small subset of the Clickbench dataset. It will run the queries for both Turso and SQLite, and print the results.

Comparing VFS's/IO Back-ends (io_uring | syscall)

make bench-vfs SQL="select * from users;" N=500

The naive script will build and run limbo in release mode and execute the given SQL (against a copy of the testing/testing.db file) N times with each vfs. This is not meant to be a definitive or thorough performance benchmark but serves to compare the two.

TPC-H

on linux if you are using tlp to manage power settings, you may want to disable it while running the TPC-H benchmark as it can affect performance. consider changing swap to swapoff -a

Run the benchmark script:

./perf/tpc-h/benchmark.sh