mirror of
https://github.com/neondatabase/neon.git
synced 2026-01-07 05:22:56 +00:00
cd982a82ec
## Problem Currently, the heap profiling frequency is every 1 MB allocated. Taking a profile stack trace takes about 1 µs, and allocating 1 MB takes about 15 µs, so the overhead is about 6.7% which is a bit high. This is a fixed cost regardless of whether heap profiles are actually accessed. ## Summary of changes Increase the heap profiling sample frequency from 1 MB to 2 MB, which reduces the overhead to about 3.3%. This seems acceptable, considering performance-sensitive code will avoid allocations as far as possible anyway.
Safekeeper Benchmarks
To run benchmarks:
# All benchmarks.
cargo bench --package safekeeper
# Specific file.
cargo bench --package safekeeper --bench receive_wal
# Specific benchmark.
cargo bench --package safekeeper --bench receive_wal process_msg/fsync=false
# List available benchmarks.
cargo bench --package safekeeper --benches -- --list
# Generate flamegraph profiles using pprof-rs, profiling for 10 seconds.
# Output in target/criterion/*/profile/flamegraph.svg.
cargo bench --package safekeeper --bench receive_wal process_msg/fsync=false --profile-time 10
Additional charts and statistics are available in target/criterion/report/index.html.
Benchmarks are automatically compared against the previous run. To compare against other runs, see
--baseline and --save-baseline.