rspack-perf-profiling

Rspack Perf Profiling

Overview

Profile Rspack builds on Linux using perf with DWARF call graphs, capture kernel + user stacks, and analyze hotspots directly from perf.data. Optionally import into samply for per-CPU threads visualization.

Workflow

1) Build profiling-enabled binding (once per code change)

pnpm run build:binding:profiling

2) Enable kernel symbols (recommended)

echo 0 | sudo tee /proc/sys/kernel/kptr_restrict
echo 1 | sudo tee /proc/sys/kernel/perf_event_paranoid

Optional: install vmlinux debug symbols or pass a vmlinux path to perf report.

3) Record perf profile (example: react-10k case)

# if benchmark repo isn't present yet (clone alongside rspack, not inside it)
git -c http.lowSpeedLimit=1 -c http.lowSpeedTime=600 \
  clone https://github.com/rstackjs/build-tools-performance.git \
  ../build-tools-performance

# install benchmark deps (required for react-10k)
pnpm -C ../build-tools-performance install

# link local rspack core so cases can resolve @rspack/core
pnpm -C ../build-tools-performance add -w @rspack/core@link:../rspack/packages/rspack

(
  cd ../build-tools-performance/cases/react-10k || exit 1
  perf record -o ./perf.data \
    -e cycles:uk -F 4000 --call-graph dwarf -- \
    node --perf-prof --perf-basic-prof --interpreted-frames-native-stack \
    ../../../rspack/packages/rspack-cli/bin/rspack.js \
    -c ./rspack.config.mjs
)

Notes:

• cycles:uk captures user + kernel cycles.
• Increase -F for higher sample density; expect large perf.data.
• Ensure --call-graph dwarf for readable Rust stacks.

4) Analyze perf.data (perf-based)

Top hotspots (flat view):

perf report -i ../build-tools-performance/cases/react-10k/perf.data \
  --stdio --no-children -g none --percent-limit 0.5 | head -n 100

Callgraph (if needed):

perf report -i ../build-tools-performance/cases/react-10k/perf.data \
  --stdio --no-children -g graph,0.5,caller,function,percent | head -n 120

5) Optional: import into samply with per-CPU threads

samply import ../build-tools-performance/cases/react-10k/perf.data \
  --per-cpu-threads -o ../build-tools-performance/cases/react-10k/perf.profile.json.gz \
  --no-open

Use this only for visualization; keep analysis perf-first.

Variants

• For other cases, swap -c <case>/rspack.config.js.
• For heavier workloads, wrap the rspack command in a loop to amplify time.
• If kernel symbols are still missing, pass -k /path/to/vmlinux to perf report.