Files
polyfill-rs/scripts/network_benchmark.sh

209 lines
5.6 KiB
Bash
Executable File

#!/bin/bash
# Network benchmark script for polyfill-rs
# Tests real network latency vs computational performance
set -e
echo "🌐 Network Latency Benchmark for polyfill-rs"
echo "============================================="
echo "📝 Note: API credentials loaded from .env file"
echo ""
# Colors for output
RED='\033[0;31m'
GREEN='\033[0;32m'
BLUE='\033[0;34m'
YELLOW='\033[1;33m'
NC='\033[0m' # No Color
print_header() {
echo -e "${BLUE}$1${NC}"
}
print_success() {
echo -e "${GREEN}$1${NC}"
}
print_warning() {
echo -e "${YELLOW}⚠️ $1${NC}"
}
print_error() {
echo -e "${RED}$1${NC}"
}
# Test network connectivity
test_connectivity() {
print_header "Testing Polymarket API connectivity..."
if curl -s --max-time 10 "https://clob.polymarket.com/ok" > /dev/null; then
print_success "Polymarket API is reachable"
return 0
else
print_error "Cannot reach Polymarket API"
return 1
fi
}
# Run network benchmarks
run_network_benchmarks() {
print_header "Running network benchmarks..."
echo "Building benchmarks..."
cargo build --release --benches
echo "Running network latency tests..."
cargo bench --bench network_benchmarks
print_success "Network benchmarks completed"
}
# Compare with computational benchmarks
run_comparison() {
print_header "Running computational vs network comparison..."
echo "1. Computational benchmarks (no network):"
cargo bench --bench comparison_benchmarks
echo ""
echo "2. Network benchmarks (with real API calls):"
cargo bench --bench network_benchmarks
print_success "Comparison completed"
}
# Manual timing test
manual_timing_test() {
print_header "Manual timing test..."
echo "Testing real API calls with manual timing:"
cat > /tmp/timing_test.rs << 'EOF'
use polyfill_rs::ClobClient;
use std::time::Instant;
#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
let client = ClobClient::new("https://clob.polymarket.com");
println!("Testing simplified markets endpoint...");
let start = Instant::now();
match client.get_sampling_simplified_markets(None).await {
Ok(markets) => {
let duration = start.elapsed();
println!("✅ Fetched {} markets in {:?}", markets.data.len(), duration);
}
Err(e) => {
let duration = start.elapsed();
println!("❌ Error after {:?}: {}", duration, e);
}
}
println!("\nTesting full markets endpoint...");
let start = Instant::now();
match client.get_sampling_markets(None).await {
Ok(markets) => {
let duration = start.elapsed();
println!("✅ Fetched {} markets in {:?}", markets.data.len(), duration);
}
Err(e) => {
let duration = start.elapsed();
println!("❌ Error after {:?}: {}", duration, e);
}
}
Ok(())
}
EOF
echo "Compiling timing test..."
rustc --edition 2021 -L target/release/deps /tmp/timing_test.rs -o /tmp/timing_test \
--extern polyfill_rs=target/release/libpolyfill_rs.rlib \
--extern tokio=target/release/deps/libtokio*.rlib 2>/dev/null || {
print_warning "Could not compile timing test. Running with cargo instead..."
# Create a simple example instead
cargo run --release --example benchmark_demo | grep -E "(Fetch|Average|Error)"
return
}
echo "Running timing test..."
/tmp/timing_test
print_success "Manual timing test completed"
}
# Generate network vs computational report
generate_network_report() {
print_header "Generating network performance report..."
cat << 'EOF'
📊 NETWORK vs COMPUTATIONAL PERFORMANCE
=======================================
To get fair benchmarks against polymarket-rs-client, we need to measure:
1. **Computational Performance** (what we currently measure):
- JSON parsing: ~2.4µs
- Order creation: ~19.7ns (struct creation only)
- Order book ops: ~118µs per 1000 updates
2. **Network Performance** (what original benchmarks measure):
- Full HTTP request + JSON parsing
- EIP-712 signing + network round-trip
- Real-world latency including server response time
3. **Fair Comparison Requirements**:
- Same network conditions (geographic location)
- Same API endpoints and request patterns
- Same authentication and signing overhead
🔬 **To run network benchmarks**:
./scripts/network_benchmark.sh
📈 **Expected results**:
- Network latency will dominate (100-500ms typical)
- Our computational improvements still provide benefits
- Total time = Network Time + Computational Time
- We optimize the computational portion significantly
🎯 **Real-world impact**:
- In co-located environments: computational speed matters more
- Over internet: network dominates, but every microsecond counts
- For high-frequency operations: our optimizations compound
EOF
print_success "Report generated"
}
# Main execution
main() {
if ! test_connectivity; then
print_error "Cannot proceed without network connectivity"
generate_network_report
exit 1
fi
# Build first
print_header "Building polyfill-rs..."
cargo build --release
# Run tests
manual_timing_test
echo ""
run_network_benchmarks
echo ""
generate_network_report
print_success "Network benchmark analysis completed!"
echo ""
echo "📁 Detailed results in target/criterion/*/report/index.html"
echo "🔗 Compare with: https://github.com/polymarket-rs-client"
}
# Run main function
main "$@"