Skip to content

fix: derive benchmark Throughput and Mean time per paragraph from ns/op #57

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
nicobistolfi merged 1 commit into from
Aug 11, 2026
Merged

fix: derive benchmark Throughput and Mean time per paragraph from ns/op #57

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
41 changes: 29 additions & 12 deletions .github/workflows/benchmark-profile.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -162,14 +162,24 @@ jobs:
const cpuTop20 = fs.readFileSync('profiles/cpu_top20.txt', 'utf8');
const memTop20 = fs.readFileSync('profiles/mem_top20.txt', 'utf8');

// Extract key metrics from benchmark output
const throughputMatch = benchmarkResults.match(/Throughput: ([\d.]+) paragraphs\/second/);
const meanTimeMatch = benchmarkResults.match(/Mean time per paragraph: ([\d.]+\w+)/);
// Extract the mean time per operation from the standard `go test -bench` output,
// then derive throughput and a human-readable mean time from it. Go's benchmark
// output never contains literal "Throughput:" or "Mean time per paragraph:" text,
// so those values must be computed rather than matched.
const nsOpMatch = benchmarkResults.match(/BenchmarkParagraphDeidentification.*?\s+([\d.]+) ns\/op/);

const throughput = throughputMatch ? throughputMatch[1] : 'N/A';
const meanTime = meanTimeMatch ? meanTimeMatch[1] : 'N/A';
const nsOp = nsOpMatch ? nsOpMatch[1] : 'N/A';
const nsOp = nsOpMatch ? parseFloat(nsOpMatch[1]) : NaN;

let throughput = 'N/A';
let meanTime = 'N/A';
if (!isNaN(nsOp) && nsOp > 0) {
throughput = (1e9 / nsOp).toFixed(2);
meanTime = nsOp >= 1e6
? `${(nsOp / 1e6).toFixed(2)} ms`
: nsOp >= 1e3
? `${(nsOp / 1e3).toFixed(2)} µs`
: `${nsOp.toFixed(2)} ns`;
}
const nsOpDisplay = nsOpMatch ? nsOpMatch[1] : 'N/A';

// Get artifact URLs
const artifactUrl = `https://github.com/${{ github.repository }}/actions/runs/${{ github.run_id }}`;
Expand All @@ -187,7 +197,7 @@ jobs:
'### Performance Summary',
`- **Throughput**: ${throughput} paragraphs/second`,
`- **Mean time per paragraph**: ${meanTime}`,
`- **Nanoseconds per operation**: ${nsOp} ns/op`,
`- **Nanoseconds per operation**: ${nsOpDisplay} ns/op`,
'',
'### 🔍 Interactive Profile Visualizations',
'',
Expand Down Expand Up @@ -269,11 +279,18 @@ jobs:
echo "## Benchmark & Profile Summary" >> $GITHUB_STEP_SUMMARY
echo "" >> $GITHUB_STEP_SUMMARY

# Extract and display key metrics
if grep -q "Throughput:" profiles/benchmark.txt; then
# Derive throughput and mean time per paragraph from the ns/op value in the
# standard `go test -bench` output (that output never contains literal
# "Throughput:" or "Mean time per paragraph:" text, so it must be computed).
NS_OP=$(awk '/BenchmarkParagraphDeidentification/ && /ns\/op/ {for (i = 1; i <= NF; i++) if ($i == "ns/op") print $(i - 1); exit}' profiles/benchmark.txt)
if [ -n "$NS_OP" ]; then
echo "### Performance Metrics" >> $GITHUB_STEP_SUMMARY
grep "Throughput:" profiles/benchmark.txt >> $GITHUB_STEP_SUMMARY
grep "Mean time per paragraph:" profiles/benchmark.txt >> $GITHUB_STEP_SUMMARY
awk -v ns="$NS_OP" 'BEGIN { printf "- **Throughput**: %.2f paragraphs/second\n", 1000000000 / ns }' >> $GITHUB_STEP_SUMMARY
awk -v ns="$NS_OP" 'BEGIN {
if (ns >= 1000000) printf "- **Mean time per paragraph**: %.2f ms\n", ns / 1000000;
else if (ns >= 1000) printf "- **Mean time per paragraph**: %.2f µs\n", ns / 1000;
else printf "- **Mean time per paragraph**: %.2f ns\n", ns;
}' >> $GITHUB_STEP_SUMMARY
echo "" >> $GITHUB_STEP_SUMMARY
fi

Expand Down
Loading