Running benchmarks

Jaeger benchmarking

The Jaeger benchmarks/tracing is wrapped by a binary. The motivation of the wrapper is to allow greater control over benchmark parameters. To run benchmarks with Jaeger, you need to do the following

Spin up Jaeger service on port 6831 to receive the benchmarks trace data, and provides Jaeger UI on port 16686.
bash
```
docker run --rm -d --name jaeger -p 6831:6831/udp -p 16686:16686 jaegertracing/all-in-one:1.62.0
```

See all the options to run a benchmark.

bash

cargo run --release --bin proof-of-sql-benches -- --help

Navigate to http://localhost:16686/ to see the results.
To end the Jaeger service, run
bash
```
docker kill jaeger
```

All the options are outlined in the help and main.rs module.

Example

To run a benchmark on the HyperKZG commitment scheme using the Single Column Filter query with a table size of 1_000_000 for 3 iterations, your command would be the following.

bash

cargo run --release --bin proof-of-sql-benches -- -s hyper-kzg -i 3 -t 1000000 -q single-column-filter

Memory logging (optional)

Jaeger benchmarks default to logging any traces at DEBUG level and above. Memory consumption is logged at TRACE level. In order to capture memory consumption in the Jaeger benchmarks, add RUST_LOG=trace to the command.

Example

RUST_LOG=trace cargo run --release --bin proof-of-sql-benches

Criterion benchmarking

To run benchmarks with Criterion, you need to do the following

Run the benchmarks. (Warning: this takes a very long time.)

bash

cargo bench -p proof-of-sql-benches --bench bench_append_rows --features="test"

Navigate to target/criterion/report/index.html to see the results.