Spaces:
Sleeping
Sleeping
File size: 11,518 Bytes
ba6114e | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 | // benchmark/bench_matching_engine.cpp
// Performance benchmark β measures throughput and latency under load
// Build: cmake --build build --target hft_bench && ./build/hft_bench
#include "core/order_book.hpp"
#include "core/simulator.hpp"
#include "core/types.hpp"
#include "journal/journal.hpp"
#include <iostream>
#include <vector>
#include <chrono>
#include <random>
#include <algorithm>
#include <iomanip>
#include <cmath>
#include <numeric>
using namespace hft;
using Clock = std::chrono::high_resolution_clock;
struct BenchResult {
std::string name;
int64_t ops;
double total_us;
double ops_per_sec;
double p50_ns;
double p99_ns;
double p999_ns;
double avg_ns;
};
static BenchResult run_bench(const std::string& name, int64_t ops,
std::vector<int64_t>& latencies) {
std::sort(latencies.begin(), latencies.end());
double total_ns = 0;
for (auto l : latencies) total_ns += l;
BenchResult r;
r.name = name;
r.ops = ops;
r.total_us = total_ns / 1000.0;
r.ops_per_sec = ops / (total_ns / 1e9);
r.avg_ns = total_ns / ops;
r.p50_ns = latencies[latencies.size() * 50 / 100];
r.p99_ns = latencies[latencies.size() * 99 / 100];
r.p999_ns = latencies[latencies.size() * 999 / 1000];
return r;
}
static void print_result(const BenchResult& r) {
std::cout << std::left << std::setw(35) << r.name
<< std::right
<< " ops: " << std::setw(8) << r.ops
<< " throughput: " << std::setw(10) << std::fixed << std::setprecision(0) << r.ops_per_sec << " ops/s"
<< " avg: " << std::setw(8) << std::setprecision(1) << r.avg_ns << " ns"
<< " p50: " << std::setw(8) << r.p50_ns << " ns"
<< " p99: " << std::setw(8) << r.p99_ns << " ns"
<< " p999: " << std::setw(8) << r.p999_ns << " ns"
<< "\n";
}
// βββ Benchmark 1: Order Insert (no match) βββββββββββββββββββββββββββββββββββ
static BenchResult bench_insert_no_match(int n) {
OrderBook book("BENCH");
std::mt19937 rng(42);
std::uniform_int_distribution<int> price_dist(1000, 2000); // spread out
std::uniform_int_distribution<int> qty_dist(100, 1000);
std::vector<int64_t> latencies;
latencies.reserve(n);
for (int i = 0; i < n; ++i) {
Order o;
o.id = i + 1;
o.symbol = "BENCH";
o.side = (i % 2 == 0) ? Side::BUY : Side::SELL;
o.type = OrderType::LIMIT;
// Spread prices so no matching occurs (buys low, sells high)
o.price = (o.side == Side::BUY) ? to_price(100.0 + price_dist(rng) * 0.01)
: to_price(200.0 + price_dist(rng) * 0.01);
o.qty = qty_dist(rng);
auto t0 = Clock::now();
book.add_order(o);
auto t1 = Clock::now();
latencies.push_back(std::chrono::duration_cast<std::chrono::nanoseconds>(t1 - t0).count());
}
return run_bench("Insert (no match)", n, latencies);
}
// βββ Benchmark 2: Order Match (aggressive orders) βββββββββββββββββββββββββββ
static BenchResult bench_match(int n) {
OrderBook book("BENCH");
std::mt19937 rng(123);
std::uniform_int_distribution<int> qty_dist(50, 200);
std::vector<int64_t> latencies;
latencies.reserve(n);
// Pre-populate the book with resting orders
for (int i = 0; i < n; ++i) {
Order o;
o.id = i + 1;
o.symbol = "BENCH";
o.side = Side::SELL;
o.type = OrderType::LIMIT;
o.price = to_price(100.0 + (i % 50) * 0.01); // 50 price levels
o.qty = qty_dist(rng);
book.add_order(o);
}
// Now send aggressive buy orders that match
for (int i = 0; i < n; ++i) {
Order o;
o.id = n + i + 1;
o.symbol = "BENCH";
o.side = Side::BUY;
o.type = OrderType::MARKET;
o.price = 0;
o.qty = 1; // small qty to match single order partially
auto t0 = Clock::now();
book.add_order(o);
auto t1 = Clock::now();
latencies.push_back(std::chrono::duration_cast<std::chrono::nanoseconds>(t1 - t0).count());
}
return run_bench("Match (market order)", n, latencies);
}
// βββ Benchmark 3: Cancel Order ββββββββββββββββββββββββββββββββββββββββββββββ
static BenchResult bench_cancel(int n) {
OrderBook book("BENCH");
std::mt19937 rng(456);
std::uniform_int_distribution<int> price_dist(1, 1000);
std::vector<int64_t> latencies;
latencies.reserve(n);
// Insert orders
for (int i = 0; i < n; ++i) {
Order o;
o.id = i + 1;
o.symbol = "BENCH";
o.side = Side::BUY;
o.type = OrderType::LIMIT;
o.price = to_price(100.0 + price_dist(rng) * 0.01);
o.qty = 100;
book.add_order(o);
}
// Cancel in random order
std::vector<int> ids(n);
std::iota(ids.begin(), ids.end(), 1);
std::shuffle(ids.begin(), ids.end(), rng);
for (int id : ids) {
auto t0 = Clock::now();
book.cancel_order(id);
auto t1 = Clock::now();
latencies.push_back(std::chrono::duration_cast<std::chrono::nanoseconds>(t1 - t0).count());
}
return run_bench("Cancel (random)", n, latencies);
}
// βββ Benchmark 4: Snapshot ββββββββββββββββββββββββββββββββββββββββββββββββββ
static BenchResult bench_snapshot(int n) {
OrderBook book("BENCH");
// Build a book with many levels
for (int i = 0; i < 500; ++i) {
Order buy, sell;
buy.id = i * 2 + 1; buy.symbol = "BENCH"; buy.side = Side::BUY;
buy.type = OrderType::LIMIT; buy.price = to_price(100.0 - i * 0.01); buy.qty = 100;
sell.id = i * 2 + 2; sell.symbol = "BENCH"; sell.side = Side::SELL;
sell.type = OrderType::LIMIT; sell.price = to_price(101.0 + i * 0.01); sell.qty = 100;
book.add_order(buy);
book.add_order(sell);
}
std::vector<int64_t> latencies;
latencies.reserve(n);
for (int i = 0; i < n; ++i) {
auto t0 = Clock::now();
auto snap = book.snapshot(10);
auto t1 = Clock::now();
latencies.push_back(std::chrono::duration_cast<std::chrono::nanoseconds>(t1 - t0).count());
(void)snap;
}
return run_bench("Snapshot (10 levels)", n, latencies);
}
// βββ Benchmark 5: Full Pipeline (submit_order via ExchangeSimulator) ββββββββ
static BenchResult bench_full_pipeline(int n) {
RiskLimits limits;
limits.max_order_qty = 100000;
limits.max_position = 10000000;
limits.max_notional_usd = 1e12;
limits.max_orders_per_sec = 1000000;
ExchangeSimulator sim(limits);
sim.add_symbol("BENCH");
std::mt19937 rng(789);
std::uniform_int_distribution<int> price_dist(9900, 10100);
std::uniform_int_distribution<int> qty_dist(1, 10);
std::vector<int64_t> latencies;
latencies.reserve(n);
for (int i = 0; i < n; ++i) {
Order o;
o.id = sim.next_order_id();
o.symbol = "BENCH";
o.side = (i % 2 == 0) ? Side::BUY : Side::SELL;
o.type = OrderType::LIMIT;
o.price = to_price(price_dist(rng) * 0.01); // ~99.00 - 101.00
o.qty = qty_dist(rng);
auto t0 = Clock::now();
sim.submit_order(o);
auto t1 = Clock::now();
latencies.push_back(std::chrono::duration_cast<std::chrono::nanoseconds>(t1 - t0).count());
}
return run_bench("Full pipeline (risk+match)", n, latencies);
}
// βββ Benchmark 6: Backtest Throughput ββββββββββββββββββββββββββββββββββββββββ
static void bench_backtest_throughput() {
std::cout << "\n--- Backtest Throughput ---\n";
int ticks = 50000;
MarketDataParser::Config cfg;
cfg.symbol = "BENCH"; cfg.start_price = 100.0;
cfg.ticks = ticks; cfg.volatility = 0.001; cfg.seed = 42;
auto tick_data = MarketDataParser::generate_synthetic(cfg);
RiskLimits rl;
rl.max_order_qty = 10000;
rl.max_position = 1000000;
rl.max_notional_usd = 1e12;
rl.max_orders_per_sec = 100000;
auto t0 = Clock::now();
Backtester bt(rl);
bt.add_symbol("BENCH");
bt.set_strategy(std::make_unique<MarketMakingStrategy>(5.0, 100));
auto result = bt.run(tick_data, false);
auto t1 = Clock::now();
double elapsed_s = std::chrono::duration<double>(t1 - t0).count();
double throughput = ticks / elapsed_s;
std::cout << " Ticks: " << ticks << "\n";
std::cout << " Elapsed: " << std::fixed << std::setprecision(3) << elapsed_s << " s\n";
std::cout << " Throughput: " << std::setprecision(0) << throughput << " ticks/s\n";
std::cout << " Orders: " << result.orders_submitted << "\n";
std::cout << " Fills: " << result.orders_filled << "\n";
std::cout << " Trades: " << result.trades_count << "\n";
std::cout << " PnL: " << std::setprecision(2) << result.total_pnl << "\n";
}
// βββ Main ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
int main() {
std::cout << "\n";
std::cout << "βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ\n";
std::cout << "β Orderflow Replay Engine β Performance Benchmark β\n";
std::cout << "βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ\n\n";
const int N = 100000;
std::vector<BenchResult> results;
results.push_back(bench_insert_no_match(N));
results.push_back(bench_match(N));
results.push_back(bench_cancel(N));
results.push_back(bench_snapshot(N));
results.push_back(bench_full_pipeline(N));
std::cout << "--- Latency Results (N = " << N << " ops each) ---\n\n";
for (auto& r : results) print_result(r);
bench_backtest_throughput();
std::cout << "\n--- Memory ---\n";
std::cout << " sizeof(Order): " << sizeof(Order) << " bytes\n";
std::cout << " sizeof(Trade): " << sizeof(Trade) << " bytes\n";
std::cout << " sizeof(PriceLevel): " << sizeof(PriceLevel) << " bytes\n";
std::cout << " sizeof(MarketDataTick): " << sizeof(MarketDataTick) << " bytes\n";
std::cout << "\nDone.\n\n";
return 0;
}
|