foxhunt/tests/e2e/tests/performance_validation_tests.rs

use std::collections::VecDeque;
use std::sync::Arc;
use tokio::time::{timeout, Duration};
use trading_engine::{
    infrastructure::{PerformanceAnalyzer, ResourceMonitor, ThroughputMonitor},
    lockfree::{AtomicCounter, RingBuffer},
    prelude::*,
    simd::SimdProcessor,
    timing::{HardwareTimestamp, LatencyTracker, PrecisionTimer},
    trading::{OrderManager, PositionManager},
    types::{PerformanceMetrics, Price, Quantity, Symbol},
};

/// Comprehensive performance validation and benchmarking tests
pub struct PerformanceValidationTests {
    latency_tracker: Arc<LatencyTracker>,
    throughput_monitor: Arc<ThroughputMonitor>,
    resource_monitor: Arc<ResourceMonitor>,
    performance_analyzer: Arc<PerformanceAnalyzer>,
    order_manager: Arc<OrderManager>,
    position_manager: Arc<PositionManager>,
    simd_processor: Arc<SimdProcessor>,
    ring_buffer: Arc<RingBuffer<u64>>,
}

impl PerformanceValidationTests {
    pub async fn new() -> Result<Self> {
        let config = load_test_config().await?;

        let latency_tracker = Arc::new(LatencyTracker::new());
        let throughput_monitor = Arc::new(ThroughputMonitor::new());
        let resource_monitor = Arc::new(ResourceMonitor::new());
        let performance_analyzer = Arc::new(PerformanceAnalyzer::new(config.clone()));
        let order_manager = Arc::new(OrderManager::new(config.clone()).await?);
        let position_manager = Arc::new(PositionManager::new(config.clone()).await?);
        let simd_processor = Arc::new(SimdProcessor::new());
        let ring_buffer = Arc::new(RingBuffer::new(65536)); // 64k entries

        Ok(Self {
            latency_tracker,
            throughput_monitor,
            resource_monitor,
            performance_analyzer,
            order_manager,
            position_manager,
            simd_processor,
            ring_buffer,
        })
    }

    /// Test 1: Critical path sub-50μs latency validation
    /// Steps: 12 comprehensive latency measurement phases
    pub async fn test_critical_path_latency_validation(&self) -> Result<WorkflowTestResult> {
        let mut result = WorkflowTestResult::new("Critical Path Latency Validation");

        // Step 1: Hardware timing calibration
        result.add_step("Hardware Timing Calibration").await;
        let calibration_start = HardwareTimestamp::now();
        let calibration_samples: Vec<u64> = (0..10000)
            .map(|_| {
                let start = HardwareTimestamp::now();
                std::hint::black_box(42); // Prevent optimization
                start.elapsed_nanos()
            })
            .collect();

        let calibration_time = calibration_start.elapsed_nanos();
        let min_resolution = calibration_samples
            .iter()
            .filter(|&&x| x > 0)
            .min()
            .unwrap_or(&1);
        let avg_resolution =
            calibration_samples.iter().sum::<u64>() as f64 / calibration_samples.len() as f64;

        assert!(
            *min_resolution <= 50,
            "Hardware timing resolution should be ≤50ns, got {}ns",
            min_resolution
        );
        result.add_metric("hardware_resolution_ns", *min_resolution as f64);
        result.add_metric("avg_resolution_ns", avg_resolution);
        result.add_metric("calibration_time_ns", calibration_time as f64);

        // Step 2: RDTSC precision measurement
        result.add_step("RDTSC Precision Measurement").await;
        let rdtsc_samples: Vec<u64> = (0..1000)
            .map(|_| {
                let start = HardwareTimestamp::rdtsc_start();
                // Minimal operation to measure
                let _dummy = 1u64.wrapping_add(2);
                HardwareTimestamp::rdtsc_end(start)
            })
            .collect();

        let rdtsc_p50 = percentile(&rdtsc_samples, 50.0);
        let rdtsc_p95 = percentile(&rdtsc_samples, 95.0);
        let rdtsc_p99 = percentile(&rdtsc_samples, 99.0);

        assert!(
            rdtsc_p95 <= 100,
            "RDTSC P95 should be ≤100ns, got {}ns",
            rdtsc_p95
        );
        result.add_metric("rdtsc_p50_ns", rdtsc_p50 as f64);
        result.add_metric("rdtsc_p95_ns", rdtsc_p95 as f64);
        result.add_metric("rdtsc_p99_ns", rdtsc_p99 as f64);

        // Step 3: Order creation latency measurement
        result.add_step("Order Creation Latency").await;
        let order_creation_samples: Vec<u64> = (0..10000)
            .map(|i| {
                let start = HardwareTimestamp::now();
                let _order = Order::new(
                    OrderId::from_u64(i as u64),
                    Symbol::new("EURUSD"),
                    OrderType::Market,
                    OrderSide::Buy,
                    Quantity::from(100_000),
                    None,
                );
                start.elapsed_nanos()
            })
            .collect();

        let creation_p50 = percentile(&order_creation_samples, 50.0);
        let creation_p95 = percentile(&order_creation_samples, 95.0);
        let creation_p99 = percentile(&order_creation_samples, 99.0);

        assert!(
            creation_p95 <= 5_000,
            "Order creation P95 should be ≤5μs, got {}ns",
            creation_p95
        );
        result.add_metric("order_creation_p50_ns", creation_p50 as f64);
        result.add_metric("order_creation_p95_ns", creation_p95 as f64);

        // Step 4: Risk validation latency
        result.add_step("Risk Validation Latency").await;
        let symbol = Symbol::new("EURUSD");
        let test_order = Order::new(
            OrderId::new(),
            symbol.clone(),
            OrderType::Market,
            OrderSide::Buy,
            Quantity::from(100_000),
            None,
        )?;

        let risk_samples: Vec<u64> = {
            let mut samples = Vec::with_capacity(1000);
            for _ in 0..1000 {
                let start = HardwareTimestamp::now();
                let _risk_result = self.order_manager.validate_risk_fast(&test_order).await;
                let elapsed = start.elapsed_nanos();
                samples.push(elapsed);
            }
            samples
        };

        let risk_p50 = percentile(&risk_samples, 50.0);
        let risk_p95 = percentile(&risk_samples, 95.0);

        assert!(
            risk_p95 <= 10_000,
            "Risk validation P95 should be ≤10μs, got {}ns",
            risk_p95
        );
        result.add_metric("risk_validation_p50_ns", risk_p50 as f64);
        result.add_metric("risk_validation_p95_ns", risk_p95 as f64);

        // Step 5: Position update latency
        result.add_step("Position Update Latency").await;
        let position_samples: Vec<u64> = {
            let mut samples = Vec::with_capacity(1000);
            for i in 0..1000 {
                let start = HardwareTimestamp::now();
                let _result = self
                    .position_manager
                    .update_position_atomic(&symbol, Quantity::from(i as i64 * 100))
                    .await;
                let elapsed = start.elapsed_nanos();
                samples.push(elapsed);
            }
            samples
        };

        let position_p50 = percentile(&position_samples, 50.0);
        let position_p95 = percentile(&position_samples, 95.0);

        assert!(
            position_p95 <= 8_000,
            "Position update P95 should be ≤8μs, got {}ns",
            position_p95
        );
        result.add_metric("position_update_p50_ns", position_p50 as f64);
        result.add_metric("position_update_p95_ns", position_p95 as f64);

        // Step 6: Lock-free data structure performance
        result.add_step("Lock-free Structure Performance").await;
        let lockfree_samples: Vec<u64> = (0..10000)
            .map(|i| {
                let start = HardwareTimestamp::now();
                let success = self.ring_buffer.try_push(i);
                let elapsed = start.elapsed_nanos();
                assert!(success, "Ring buffer push should succeed");
                elapsed
            })
            .collect();

        let lockfree_p50 = percentile(&lockfree_samples, 50.0);
        let lockfree_p95 = percentile(&lockfree_samples, 95.0);

        assert!(
            lockfree_p95 <= 500,
            "Lock-free push P95 should be ≤500ns, got {}ns",
            lockfree_p95
        );
        result.add_metric("lockfree_push_p50_ns", lockfree_p50 as f64);
        result.add_metric("lockfree_push_p95_ns", lockfree_p95 as f64);

        // Step 7: SIMD operation performance
        result.add_step("SIMD Operation Performance").await;
        let simd_data: Vec<f64> = (0..1000).map(|i| i as f64 * 1.1).collect();
        let simd_samples: Vec<u64> = (0..1000)
            .map(|_| {
                let start = HardwareTimestamp::now();
                let _result = self.simd_processor.vectorized_multiply(&simd_data, 2.0);
                start.elapsed_nanos()
            })
            .collect();

        let simd_p50 = percentile(&simd_samples, 50.0);
        let simd_p95 = percentile(&simd_samples, 95.0);

        assert!(
            simd_p95 <= 2_000,
            "SIMD operation P95 should be ≤2μs, got {}ns",
            simd_p95
        );
        result.add_metric("simd_operation_p50_ns", simd_p50 as f64);
        result.add_metric("simd_operation_p95_ns", simd_p95 as f64);

        // Step 8: End-to-end critical path measurement
        result.add_step("End-to-End Critical Path").await;
        let e2e_samples: Vec<u64> = {
            let mut samples = Vec::with_capacity(1000);
            for i in 0..1000 {
                let start = HardwareTimestamp::now();

                // Critical path: Order creation -> Risk check -> Position update -> Submit
                let order = Order::new(
                    OrderId::from_u64(10000 + i as u64),
                    symbol.clone(),
                    OrderType::Market,
                    OrderSide::Buy,
                    Quantity::from(100_000),
                    None,
                )?;

                let _risk_check = self.order_manager.validate_risk_fast(&order).await;
                let _position_update = self
                    .position_manager
                    .update_position_atomic(&symbol, Quantity::from(100_000))
                    .await;
                let _submission = self.order_manager.submit_order_fast(order).await;

                let elapsed = start.elapsed_nanos();
                samples.push(elapsed);
            }
            Result::<Vec<u64>>::Ok(samples)
        }?;

        let e2e_p50 = percentile(&e2e_samples, 50.0);
        let e2e_p95 = percentile(&e2e_samples, 95.0);
        let e2e_p99 = percentile(&e2e_samples, 99.0);

        // THE CRITICAL REQUIREMENT: Sub-50μs end-to-end
        assert!(
            e2e_p95 < 50_000,
            "End-to-end critical path too slow: P95={}ns > 50μs",
            e2e_p95
        );
        result.add_metric("e2e_critical_p50_ns", e2e_p50 as f64);
        result.add_metric("e2e_critical_p95_ns", e2e_p95 as f64);
        result.add_metric("e2e_critical_p99_ns", e2e_p99 as f64);

        // Step 9: Jitter analysis
        result.add_step("Jitter Analysis").await;
        let jitter_samples: Vec<u64> = e2e_samples
            .windows(2)
            .map(|pair| (pair[1] as i64 - pair[0] as i64).abs() as u64)
            .collect();

        let jitter_p95 = percentile(&jitter_samples, 95.0);
        let jitter_max = jitter_samples.iter().max().unwrap_or(&0);

        assert!(
            jitter_p95 < 10_000,
            "Jitter P95 should be <10μs, got {}ns",
            jitter_p95
        );
        result.add_metric("jitter_p95_ns", jitter_p95 as f64);
        result.add_metric("jitter_max_ns", *jitter_max as f64);

        // Step 10: Temperature and throttling monitoring
        result.add_step("Thermal Performance").await;
        let thermal_metrics = self.resource_monitor.get_thermal_metrics().await?;
        assert!(
            thermal_metrics.cpu_temperature_celsius < 80.0,
            "CPU temperature too high: {}°C",
            thermal_metrics.cpu_temperature_celsius
        );
        assert!(
            !thermal_metrics.is_throttling,
            "CPU should not be throttling"
        );
        result.add_metric("cpu_temperature", thermal_metrics.cpu_temperature_celsius);

        // Step 11: Cache performance analysis
        result.add_step("Cache Performance Analysis").await;
        let cache_metrics = self.resource_monitor.get_cache_metrics().await?;
        assert!(
            cache_metrics.l1_hit_rate > 0.95,
            "L1 cache hit rate should be >95%"
        );
        assert!(
            cache_metrics.l2_hit_rate > 0.90,
            "L2 cache hit rate should be >90%"
        );
        result.add_metric("l1_hit_rate", cache_metrics.l1_hit_rate);
        result.add_metric("l2_hit_rate", cache_metrics.l2_hit_rate);

        // Step 12: Sustained performance validation
        result.add_step("Sustained Performance").await;
        let sustained_start = HardwareTimestamp::now();
        let mut sustained_samples = Vec::with_capacity(10000);

        // Run for 10 seconds at high frequency
        let test_duration = Duration::from_secs(10);
        let test_end = sustained_start.add_duration(test_duration);

        while HardwareTimestamp::now() < test_end {
            let sample_start = HardwareTimestamp::now();

            let order = Order::new(
                OrderId::new(),
                symbol.clone(),
                OrderType::Market,
                OrderSide::Buy,
                Quantity::from(100_000),
                None,
            )?;

            let _risk_check = self.order_manager.validate_risk_fast(&order).await;
            let elapsed = sample_start.elapsed_nanos();
            sustained_samples.push(elapsed);
        }

        let sustained_p95 = percentile(&sustained_samples, 95.0);
        let sustained_degradation = (sustained_p95 as f64 / e2e_p95 as f64) - 1.0;

        assert!(
            sustained_degradation < 0.20,
            "Sustained performance degradation should be <20%, got {:.1}%",
            sustained_degradation * 100.0
        );
        result.add_metric("sustained_samples", sustained_samples.len() as f64);
        result.add_metric("sustained_p95_ns", sustained_p95 as f64);
        result.add_metric("performance_degradation", sustained_degradation);

        result.mark_success();
        Ok(result)
    }

    /// Test 2: Throughput and scalability benchmarks
    /// Steps: 10 comprehensive throughput measurement phases
    pub async fn test_throughput_scalability_benchmarks(&self) -> Result<WorkflowTestResult> {
        let mut result = WorkflowTestResult::new("Throughput Scalability Benchmarks");

        // Step 1: Single-threaded baseline throughput
        result.add_step("Single-threaded Baseline").await;
        let single_thread_start = HardwareTimestamp::now();
        let mut operations_completed = 0u64;
        let test_duration = Duration::from_secs(5);
        let end_time = single_thread_start.add_duration(test_duration);

        while HardwareTimestamp::now() < end_time {
            let order = Order::new(
                OrderId::from_u64(operations_completed),
                Symbol::new("EURUSD"),
                OrderType::Market,
                OrderSide::Buy,
                Quantity::from(100_000),
                None,
            )?;

            let _validation = self.order_manager.validate_risk_fast(&order).await;
            operations_completed += 1;
        }

        let actual_duration = single_thread_start.elapsed_nanos() as f64 / 1_000_000_000.0;
        let single_thread_ops_per_sec = operations_completed as f64 / actual_duration;

        assert!(
            single_thread_ops_per_sec > 100_000.0,
            "Single-thread should exceed 100k ops/sec, got {:.0}",
            single_thread_ops_per_sec
        );
        result.add_metric("single_thread_ops_per_sec", single_thread_ops_per_sec);
        result.add_metric("single_thread_total_ops", operations_completed as f64);

        // Step 2: Multi-threaded throughput scaling
        result.add_step("Multi-threaded Scaling").await;
        let thread_counts = vec![2, 4, 8, 16];
        let mut scaling_results = Vec::new();

        for thread_count in thread_counts {
            let mt_start = HardwareTimestamp::now();
            let operations_per_thread = Arc::new(AtomicCounter::new());

            let handles: Vec<_> = (0..thread_count)
                .map(|thread_id| {
                    let counter = operations_per_thread.clone();
                    let order_manager = self.order_manager.clone();

                    tokio::spawn(async move {
                        let thread_start = HardwareTimestamp::now();
                        let thread_duration = Duration::from_secs(3);
                        let thread_end = thread_start.add_duration(thread_duration);

                        let mut local_ops = 0u64;
                        while HardwareTimestamp::now() < thread_end {
                            let order = Order::new(
                                OrderId::from_u64((thread_id as u64) << 32 | local_ops),
                                Symbol::new("EURUSD"),
                                OrderType::Market,
                                OrderSide::Buy,
                                Quantity::from(100_000),
                                None,
                            )?;

                            let _validation = order_manager.validate_risk_fast(&order).await;
                            local_ops += 1;
                        }

                        counter.add(local_ops);
                        Result::<u64>::Ok(local_ops)
                    })
                })
                .collect();

            let thread_results = futures::future::join_all(handles).await;
            let mt_duration = mt_start.elapsed_nanos() as f64 / 1_000_000_000.0;
            let total_mt_ops = operations_per_thread.get();
            let mt_ops_per_sec = total_mt_ops as f64 / mt_duration;

            scaling_results.push((thread_count, mt_ops_per_sec));

            // Validate scaling efficiency
            let scaling_efficiency =
                mt_ops_per_sec / (single_thread_ops_per_sec * thread_count as f64);
            result.add_metric(
                &format!("mt_{}_threads_ops_per_sec", thread_count),
                mt_ops_per_sec,
            );
            result.add_metric(
                &format!("mt_{}_threads_efficiency", thread_count),
                scaling_efficiency,
            );

            // Should maintain at least 70% efficiency up to 8 threads
            if thread_count <= 8 {
                assert!(
                    scaling_efficiency > 0.70,
                    "Scaling efficiency for {} threads too low: {:.1}%",
                    thread_count,
                    scaling_efficiency * 100.0
                );
            }
        }

        // Step 3: Memory bandwidth saturation test
        result.add_step("Memory Bandwidth Saturation").await;
        let memory_test_data: Vec<f64> = (0..1_000_000).map(|i| i as f64 * 1.1).collect();
        let memory_start = HardwareTimestamp::now();

        let memory_operations = 1000;
        for _ in 0..memory_operations {
            let _result = self.simd_processor.vectorized_sum(&memory_test_data);
        }

        let memory_duration = memory_start.elapsed_nanos() as f64 / 1_000_000.0; // ms
        let memory_bandwidth_gbps = (memory_test_data.len() * 8 * memory_operations) as f64
            / 1_000_000_000.0
            / (memory_duration / 1000.0);

        result.add_metric("memory_bandwidth_gbps", memory_bandwidth_gbps);
        assert!(
            memory_bandwidth_gbps > 10.0,
            "Memory bandwidth should exceed 10 GB/s"
        );

        // Step 4: Queue depth and batching optimization
        result.add_step("Queue Depth Optimization").await;
        let batch_sizes = vec![1, 8, 32, 128, 512];
        let mut batch_results = Vec::new();

        for batch_size in batch_sizes {
            let batch_start = HardwareTimestamp::now();
            let total_batches = 1000;

            for batch_idx in 0..total_batches {
                let mut batch_orders = Vec::with_capacity(batch_size);

                for i in 0..batch_size {
                    let order = Order::new(
                        OrderId::from_u64((batch_idx * batch_size + i) as u64),
                        Symbol::new("EURUSD"),
                        OrderType::Market,
                        OrderSide::Buy,
                        Quantity::from(100_000),
                        None,
                    )?;
                    batch_orders.push(order);
                }

                let _batch_result = self.order_manager.validate_risk_batch(&batch_orders).await;
            }

            let batch_duration = batch_start.elapsed_nanos() as f64 / 1_000_000_000.0;
            let batch_ops_per_sec = (total_batches * batch_size) as f64 / batch_duration;

            batch_results.push((batch_size, batch_ops_per_sec));
            result.add_metric(
                &format!("batch_size_{}_ops_per_sec", batch_size),
                batch_ops_per_sec,
            );
        }

        // Find optimal batch size
        let optimal_batch = batch_results
            .iter()
            .max_by(|a, b| a.1.partial_cmp(&b.1).unwrap())
            .unwrap();
        result.add_metric("optimal_batch_size", optimal_batch.0 as f64);
        result.add_metric("optimal_batch_ops_per_sec", optimal_batch.1);

        // Step 5: Network I/O throughput simulation
        result.add_step("Network I/O Throughput").await;
        let network_start = HardwareTimestamp::now();
        let message_count = 100_000;
        let message_size = 256; // bytes

        for i in 0..message_count {
            let message = vec![0u8; message_size];
            let _serialized = self.order_manager.serialize_order_message(&message).await;
        }

        let network_duration = network_start.elapsed_nanos() as f64 / 1_000_000_000.0;
        let network_messages_per_sec = message_count as f64 / network_duration;
        let network_mbps = (message_count * message_size) as f64 / 1_000_000.0 / network_duration;

        result.add_metric("network_messages_per_sec", network_messages_per_sec);
        result.add_metric("network_throughput_mbps", network_mbps);
        assert!(
            network_messages_per_sec > 50_000.0,
            "Network message rate should exceed 50k/sec"
        );

        // Step 6: Database write throughput
        result.add_step("Database Write Throughput").await;
        let db_start = HardwareTimestamp::now();
        let db_writes = 10_000;

        for i in 0..db_writes {
            let trade_record = create_test_trade_record(i);
            let _db_result = self.order_manager.persist_trade_record(&trade_record).await;
        }

        let db_duration = db_start.elapsed_nanos() as f64 / 1_000_000_000.0;
        let db_writes_per_sec = db_writes as f64 / db_duration;

        result.add_metric("db_writes_per_sec", db_writes_per_sec);
        assert!(
            db_writes_per_sec > 5_000.0,
            "Database writes should exceed 5k/sec"
        );

        // Step 7: CPU utilization under load
        result.add_step("CPU Utilization Analysis").await;
        let cpu_start = HardwareTimestamp::now();
        let baseline_cpu = self.resource_monitor.get_cpu_usage().await?;

        // Generate high load
        let high_load_duration = Duration::from_secs(5);
        let load_end = cpu_start.add_duration(high_load_duration);

        let _load_task = tokio::spawn(async move {
            while HardwareTimestamp::now() < load_end {
                // Simulate trading workload
                let _computation = (0..1000).map(|x| x * x).sum::<i32>();
            }
        });

        tokio::time::sleep(Duration::from_secs(2)).await;
        let load_cpu = self.resource_monitor.get_cpu_usage().await?;

        let cpu_utilization = load_cpu.user_percent + load_cpu.system_percent;
        result.add_metric("cpu_utilization_percent", cpu_utilization);
        result.add_metric("cpu_user_percent", load_cpu.user_percent);
        result.add_metric("cpu_system_percent", load_cpu.system_percent);

        assert!(
            cpu_utilization < 90.0,
            "CPU utilization should stay below 90%"
        );

        // Step 8: Memory allocation and GC pressure
        result.add_step("Memory Allocation Analysis").await;
        let memory_start = self.resource_monitor.get_memory_usage().await?;

        // Allocate and deallocate memory to test pressure
        let allocation_cycles = 1000;
        for _ in 0..allocation_cycles {
            let large_allocation: Vec<u64> = (0..10_000).collect();
            std::hint::black_box(&large_allocation); // Prevent optimization
        }

        let memory_end = self.resource_monitor.get_memory_usage().await?;
        let memory_growth = memory_end.used_mb - memory_start.used_mb;

        result.add_metric("memory_growth_mb", memory_growth);
        result.add_metric("memory_utilization_percent", memory_end.utilization_percent);

        // Memory growth should be reasonable (not a major leak)
        assert!(
            memory_growth < 100.0,
            "Memory growth should be <100MB for test workload"
        );

        // Step 9: I/O wait and disk performance
        result.add_step("I/O Performance Analysis").await;
        let io_metrics = self.resource_monitor.get_io_metrics().await?;
        result.add_metric("disk_read_mbps", io_metrics.read_mbps);
        result.add_metric("disk_write_mbps", io_metrics.write_mbps);
        result.add_metric("io_wait_percent", io_metrics.io_wait_percent);

        assert!(io_metrics.io_wait_percent < 20.0, "I/O wait should be <20%");

        // Step 10: Overall system performance score
        result.add_step("System Performance Score").await;
        let perf_score = self
            .performance_analyzer
            .calculate_overall_score(
                single_thread_ops_per_sec,
                optimal_batch.1,
                network_messages_per_sec,
                db_writes_per_sec,
            )
            .await?;

        result.add_metric("overall_performance_score", perf_score.total_score);
        result.add_metric("latency_score", perf_score.latency_score);
        result.add_metric("throughput_score", perf_score.throughput_score);
        result.add_metric(
            "resource_efficiency_score",
            perf_score.resource_efficiency_score,
        );

        assert!(
            perf_score.total_score > 85.0,
            "Overall performance score should exceed 85/100"
        );

        result.mark_success();
        Ok(result)
    }
}

// Helper functions for performance testing
fn percentile(samples: &[u64], percentile: f64) -> u64 {
    if samples.is_empty() {
        return 0;
    }

    let mut sorted = samples.to_vec();
    sorted.sort_unstable();

    let index = ((percentile / 100.0) * (sorted.len() - 1) as f64) as usize;
    sorted[index]
}

fn create_test_trade_record(id: u64) -> TradeRecord {
    TradeRecord {
        trade_id: TradeId::from_u64(id),
        symbol: Symbol::new("EURUSD"),
        quantity: Quantity::from(100_000),
        price: Price::from_f64(1.1050).unwrap(),
        side: OrderSide::Buy,
        timestamp: HardwareTimestamp::now(),
        venue: "TEST_VENUE".to_string(),
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[tokio::test]
    async fn test_critical_path_latency_integration() {
        let test_suite = PerformanceValidationTests::new().await.unwrap();
        let result = test_suite
            .test_critical_path_latency_validation()
            .await
            .unwrap();
        assert!(result.success, "Critical path latency test failed");
        assert!(result.steps.len() == 12, "Should have 12 steps");

        // Verify critical performance requirements
        let e2e_p95 = result.metrics.get("e2e_critical_p95_ns").unwrap();
        assert!(
            *e2e_p95 < 50_000.0,
            "End-to-end P95 latency requirement failed"
        );
    }

    #[tokio::test]
    async fn test_throughput_benchmarks_integration() {
        let test_suite = PerformanceValidationTests::new().await.unwrap();
        let result = test_suite
            .test_throughput_scalability_benchmarks()
            .await
            .unwrap();
        assert!(result.success, "Throughput benchmarks test failed");
        assert!(result.steps.len() == 10, "Should have 10 steps");

        // Verify throughput requirements
        let single_thread_ops = result.metrics.get("single_thread_ops_per_sec").unwrap();
        assert!(
            *single_thread_ops > 100_000.0,
            "Single-thread throughput requirement failed"
        );
    }
}