HammerEngine Benchmark Report Generator

This Skill creates comprehensive, professional-quality performance reports from SDL3 HammerEngine benchmark results. Perfect for documentation, stakeholder updates, or performance tracking over time.

Report Capabilities

•Data Aggregation - Collect metrics from all benchmark test results
•Statistical Analysis - Calculate mean, median, stddev, percentiles
•Visual Representation - Generate ASCII charts and tables
•Trend Analysis - Compare against historical baselines
•Callgrind Integration - Include profiling hotspot analysis
•Professional Formatting - Markdown/HTML/PDF output formats
•Executive Summary - High-level overview for stakeholders

Input Sources

Benchmark Result Files

Location: $PROJECT_ROOT/test_results/ (Ensure PROJECT_ROOT environment variable is set to project root directory)

Files to Scan:

code

$PROJECT_ROOT/test_results/
├── ai_benchmark/
│   ├── performance_metrics.txt
│   └── performance_report.md
├── collision_benchmark/
│   ├── performance_metrics.txt
│   └── performance_report.md
├── pathfinder_benchmark/
│   ├── performance_metrics.txt
│   └── performance_report.md
├── event_manager_scaling/
│   ├── performance_metrics.txt
│   └── performance_report.md
├── particle_benchmark/
│   ├── performance_metrics.txt
│   └── performance_report.md
├── ui_stress/
│   ├── performance_metrics.txt
│   └── performance_report.md
└── valgrind/
    ├── callgrind/
    │   └── callgrind.out.*
    └── cache/
        └── cachegrind.out.*

Historical Baselines (if available)

code

$PROJECT_ROOT/test_results/baseline/
$PROJECT_ROOT/test_results/baseline_history/

Execution Steps

Step 1: Discover Available Benchmark Results

Scan for results:

bash

find "$PROJECT_ROOT/test_results/" -name "performance_metrics.txt" -type f

Expected benchmarks:

•AI System Benchmark
•Collision System Benchmark
•Pathfinder Benchmark
•Event Manager Scaling
•Particle Manager Benchmark
•UI Stress Tests

For each found result:

•Record timestamp (from file modification date)
•Extract all metrics
•Categorize by system

Step 2: Extract and Parse Metrics

Extraction Patterns:

AI System Metrics (Dual Benchmark System)

Synthetic Benchmarks (AIManager infrastructure):

bash

grep -B 5 -A 10 "TestSynthetic" "$PROJECT_ROOT/test_results/ai_scaling_benchmark_"*.txt | \
  grep -E "Entity updates per second:|entities"

Integrated Benchmarks (Production behaviors):

bash

grep -B 5 -A 10 "TestIntegrated" "$PROJECT_ROOT/test_results/ai_scaling_benchmark_"*.txt | \
  grep -E "Entity updates per second:|entities"

Parse into structure:

code

AI_System_Synthetic:
  Entity_100: 170000
  Entity_200: 750000
  Entity_1000: 975000
  Entity_5000: 925000
  Entity_10000: 995000

AI_System_Integrated:
  Entity_100: 569152
  Entity_200: 579794
  Entity_500: 611098
  Entity_1000: 1192606
  Entity_2000: 1587491

Collision System Metrics

bash

grep -E "^(Collision Checks|Query Time|Hash Efficiency|AABB Tests):" \
  "$PROJECT_ROOT/test_results/collision_benchmark/performance_metrics.txt"

Pathfinder Metrics

bash

grep -E "^(Path Calculation|Nodes Explored|Cache Hits|A\* Performance):" \
  "$PROJECT_ROOT/test_results/pathfinder_benchmark/performance_metrics.txt"

Event Manager Metrics

bash

grep -E "^(Events/sec|Dispatch Latency|Queue Depth|Peak Throughput):" \
  "$PROJECT_ROOT/test_results/event_manager_scaling/performance_metrics.txt"

Particle Manager Metrics

bash

grep -E "^(Particles/frame|Render Time|Batch Count|Culling Efficiency):" \
  "$PROJECT_ROOT/test_results/particle_benchmark/performance_metrics.txt"

UI System Metrics

bash

grep -E "^(Components|Render Time|Event Handling|DPI Scaling):" \
  "$PROJECT_ROOT/test_results/ui_stress/performance_metrics.txt"

Step 3: Statistical Analysis

For each metric with multiple data points:

Calculate Statistics

python

def calculate_statistics(values):
    """Calculate statistical measures for a set of values"""
    n = len(values)
    mean = sum(values) / n

    # Median
    sorted_values = sorted(values)
    median = sorted_values[n//2] if n % 2 == 1 else \
             (sorted_values[n//2-1] + sorted_values[n//2]) / 2

    # Standard Deviation
    variance = sum((x - mean) ** 2 for x in values) / n
    stddev = variance ** 0.5

    # Percentiles
    p50 = median
    p95 = sorted_values[int(n * 0.95)]
    p99 = sorted_values[int(n * 0.99)]

    # Min/Max
    min_val = min(values)
    max_val = max(values)

    return {
        'mean': mean,
        'median': median,
        'stddev': stddev,
        'p50': p50,
        'p95': p95,
        'p99': p99,
        'min': min_val,
        'max': max_val,
        'coefficient_of_variation': (stddev / mean) * 100 if mean > 0 else 0
    }

Apply to metrics:

•If multiple benchmark runs exist, aggregate data
•Calculate statistics for each metric
•Identify outliers (values > 2 stddev from mean)

Step 4: Callgrind Analysis (if available)

Parse Callgrind Output:

bash

# Find most recent callgrind output
CALLGRIND_FILE=$(ls -t "$PROJECT_ROOT/test_results/valgrind/callgrind/callgrind.out."* 2>/dev/null | head -n 1)

if [ -f "$CALLGRIND_FILE" ]; then
    # Extract top hotspots using callgrind_annotate
    callgrind_annotate --auto=yes "$CALLGRIND_FILE" | head -n 50
fi

Extract:

•Top 10 functions by time (Ir - instruction reads)
•Top 10 functions by calls
•Call graphs for critical paths (AI update, collision detection, rendering)

Example Hotspot Data:

code

Function                               Ir        %
AIManager::updateBehaviors()           15,234M   45.2%
CollisionManager::detectCollisions()    6,123M   18.1%
PathfinderManager::calculatePath()      4,056M   12.0%

Step 5: Trend Analysis (if baselines available)

Compare Against Baseline:

For each metric:

•Load baseline value (if exists)
•Calculate percentage change
•Determine trend direction (improving/degrading/stable)
•Identify long-term trends (if multiple historical baselines)

Trend Classification:

•📈 Improving: >5% better than baseline
•📉 Degrading: >5% worse than baseline
•➡️ Stable: Within ±5% of baseline
•🆕 New: No baseline for comparison

Step 6: Generate Report Sections

Section 1: Executive Summary

markdown

# SDL3 HammerEngine Performance Report

**Generated:** YYYY-MM-DD HH:MM:SS
**Benchmark Suite Version:** <git-commit-hash>
**Branch:** <current-branch>
**Platform:** <OS-version>
**Build Type:** Debug/Release

---

## Executive Summary

### Overall Performance: ✓ EXCELLENT / ✓ GOOD / ⚠ FAIR / ❌ NEEDS IMPROVEMENT

SDL3 HammerEngine demonstrates <strong/adequate/weak> performance across all critical systems:

- **AI System:** Handles 10,000+ entities at 62 FPS with 5.8% CPU (Target: 60+ FPS, <6% CPU) ✓
- **Collision System:** 125,000 collision checks/sec, 0.08ms query time ✓
- **Pathfinding:** 8.5ms path calculation, 78% cache hit rate ✓
- **Event System:** 8,500 events/sec throughput, 0.12ms dispatch latency ✓
- **Particle System:** 5,000 particles/frame, 3.2ms render time ✓
- **UI System:** 1,000 components, 4.5ms render time ✓

### Key Achievements

- AI system exceeds 10K entity target with headroom
- Collision system optimization improved query speed by 12%
- Event throughput supports 8K+ concurrent events
- All systems meet or exceed performance targets

### Areas for Improvement

- Event dispatch latency increased 8% (monitor)
- AI behavior updates showing slight variance (9% stddev)

### Recommendation

**Status:** Ready for production
**Next Milestones:** Optimize event dispatching, stabilize AI behavior timing

Section 2: Detailed System Analysis

markdown

## Detailed Performance Analysis

### AI System - Synthetic Benchmarks (Infrastructure Performance)

#### Purpose
Tests AIManager infrastructure without integration overhead

#### Benchmark Results

| Entity Count | Value (updates/sec) | Status | Baseline | Change |
|--------------|---------------------|--------|----------|--------|
| 100 | 170K | ✓ | 170K | 0.0% |
| 200 | 750K | ✓ | 750K | 0.0% |
| 1000 | 975K | ✓ | 975K | 0.0% |
| 5000 | 925K | ✓ | 925K | 0.0% |
| 10000 | 995K | ✓ | 995K | 0.0% |

#### Threading Efficiency
- Single-threaded (100): 170K updates/sec
- Multi-threaded (5000): 925K updates/sec
- Speedup: 5.4x

#### Statistical Summary
- Mean: 963K updates/sec (across entity counts)
- Std Dev: 138K (14% CoV)
- Consistent performance across entity scales

---

### AI System - Integrated Benchmarks (Production Workload)

#### Purpose
Tests AIManager with PathfinderManager/CollisionManager integration

#### Benchmark Results

| Entity Count | Value (updates/sec) | Status | Baseline | Change |
|--------------|---------------------|--------|----------|--------|
| 100 | 569K | ✓ | 569K | 0.0% |
| 200 | 580K | ✓ | 580K | 0.0% |
| 500 | 611K | ✓ | 611K | 0.0% |
| 1000 | 1193K | ✓ | 1193K | 0.0% |
| 2000 | 1587K | ✓ | 1587K | 0.0% |

#### Threading Efficiency
- Single-threaded (100): 569K updates/sec
- Multi-threaded (2000): 1587K updates/sec
- Speedup: 2.8x

#### Statistical Summary
- Mean: 908K updates/sec (across entity counts)
- Std Dev: 444K (49% CoV)
- Performance scales with entity count

---

### AI System - Integration Overhead Analysis

#### Overhead Metrics

| Entity Count | Synthetic | Integrated | Overhead | Assessment |
|--------------|-----------|------------|----------|------------|
| 100 | 170K/s | 569K/s | -70% | Data inconsistency* |
| 200 | 750K/s | 580K/s | +23% | Expected |
| 1000 | 975K/s | 1193K/s | -22% | Data inconsistency* |
| 2000 | N/A | 1587K/s | N/A | N/A |

*Note: Negative overhead indicates synthetic values are estimates while integrated are measured.
Expected overhead: 20-40% (integrated slower due to PathfinderManager)

#### Overhead Sources
- PathfinderManager: Path requests, cache lookups, A* computation
- CollisionManager: Spatial hash queries for neighbors
- Production behaviors: Complex state machines and calculations

**Stability Analysis:**
- ✓ FPS variance low (2.9% CoV) - excellent stability
- ✓ CPU usage consistent (0.3% stddev)
- ✓ Update time predictable (<1ms variance)

#### Performance Profile (Callgrind Hotspots)

Top functions by instruction reads:
1. `AIManager::updateBehaviors()` - 45.2% (expected, main update loop)
2. `AIManager::processBatch()` - 12.3% (batch processing)
3. `BehaviorCache::lookup()` - 8.7% (cache lookups)
4. `ChaseBehavior::executeLogic()` - 6.1% (behavior logic)
5. `PathfinderManager::requestPath()` - 4.2% (pathfinding integration)

**Analysis:** Hotspot distribution is as expected. Most time in core update loop.

#### Trend Analysis

Performance trend over time (vs historical baselines):

FPS History: Jan 2025: 62.3 FPS ← Current Dec 2024: 61.8 FPS (+0.8% improvement) Nov 2024: 59.2 FPS (+5.2% improvement)

code


📈 **Trend:** Improving steadily

#### Recommendations

1. **Maintain current performance** - AI system exceeds targets
2. **Monitor behavior update variance** - Consider additional caching
3. **Document batch processing optimization** - Apply pattern to other systems

Repeat similar detailed analysis for:

•Collision System
•Pathfinding System
•Event Manager
•Particle Manager
•UI System

Section 3: Cross-System Analysis

markdown

## Cross-System Performance Comparison

### Frame Budget Analysis (60 FPS = 16.67ms budget)

| System | Time (ms) | % Budget | Status |
|--------|-----------|----------|--------|
| AI Update | 12.4 | 74.4% | ✓ |
| Collision Detection | 2.8 | 16.8% | ✓ |
| Pathfinding | 1.2 | 7.2% | ✓ |
| Event Processing | 0.5 | 3.0% | ✓ |
| Particle Update | 0.8 | 4.8% | ✓ |
| UI Rendering | 4.5 | 27.0% | ✓ |
| **Total** | **22.2** | **133.3%** | ⚠ |

**Note:** Total exceeds 100% because systems run on separate threads (update vs render).

**Update Thread Budget (60 FPS = 16.67ms):**
- AI: 12.4ms
- Collision: 2.8ms
- Pathfinding: 1.2ms
- Event: 0.5ms
- **Total Update:** 16.9ms (101% of budget) ⚠ Slight overrun

**Render Thread Budget:**
- Particle Render: 0.8ms
- UI Render: 4.5ms
- World Render: 3.2ms
- **Total Render:** 8.5ms (51% of budget) ✓ Plenty of headroom

### System Interaction Analysis

**AI ↔ Pathfinding:**
- Pathfinding requests/frame: 15
- Average latency: 8.5ms
- Cache hit rate: 78%
- ✓ Integration efficient

**Collision ↔ Pathfinding:**
- Dynamic obstacle updates: 50/frame
- Pathfinding weight adjustments: 12/frame
- ✓ Integration smooth

**Event ↔ All Systems:**
- Event throughput: 8,500 events/sec
- Dispatch latency: 0.12ms
- Queue depth: 128 events
- ✓ No bottlenecks detected

### Resource Usage Summary

**CPU Usage by System:**

AI Manager: 5.8% Collision Manager: 2.3% Pathfinder: 1.2% Event Manager: 0.8% Particle Manager: 1.5% UI Manager: 0.9% Total Engine: 12.5%

code


**Memory Usage:**

AI Manager: 45 MB Collision Manager: 32 MB Pathfinder: 18 MB Event Manager: 8 MB Particle Manager: 12 MB UI Manager: 15 MB Total Engine: 130 MB

code

Section 4: Optimization Opportunities

markdown

## Optimization Opportunities

### High Priority

1. **Update Thread Frame Budget Overrun**
   - Current: 16.9ms (101% of 16.67ms budget)
   - Impact: Occasional frame drops
   - Recommendation: Reduce AI update time by 0.5ms
   - Approach: Increase batch size or optimize behavior cache

2. **Event Dispatch Latency Increase**
   - Current: 0.12ms (up 8% from baseline)
   - Impact: Slight event processing delay
   - Recommendation: Profile event dispatch path
   - Approach: Reduce lock contention or optimize event routing

### Medium Priority

3. **Pathfinding Cache Hit Rate**
   - Current: 78%
   - Target: 85%+
   - Recommendation: Increase cache size or improve eviction policy
   - Expected Improvement: Reduce path calculation time by ~15%

4. **Particle Culling Efficiency**
   - Current: 88%
   - Target: 95%+
   - Recommendation: Improve camera frustum culling
   - Expected Improvement: Reduce render time by ~10%

### Low Priority

5. **UI Component Render Time**
   - Current: 4.5ms (stable, within budget)
   - Opportunity: Apply batching technique from particle system
   - Expected Improvement: Reduce to 3.5ms (~22% faster)

Section 5: Historical Trends

markdown

## Performance Trends

### AI System FPS Over Time

68 ┤ 66 ┤ ╭─ 64 ┤ ╭─╯ 62 ┤ ╭───╯ ← Current (62.3) 60 ┤ ╭─╯ 58 ┤╭─╯ 56 ┼───────────────────────── Nov Dec Jan 2024 2024 2025

code


**Trend:** 📈 Improving (+5.2% over 3 months)

### Collision System Query Time

0.12 ┤╮ 0.10 ┤ ╰╮ 0.08 ┤ ╰───────╮ ← Current (0.08ms) 0.06 ┤ ╰─ 0.04 ┼───────────────────────── Nov Dec Jan 2024 2024 2025

code


**Trend:** 📈 Improving (-33% over 3 months)

### Event Throughput

9000 ┤ ╭─── ← Current (8500/sec) 8500 ┤ ╭─╯ 8000 ┤ ╭─╯ 7500 ┤╭─╯ 7000 ┼───────────────────────── Nov Dec Jan 2024 2024 2025

code


**Trend:** 📈 Improving (+21% over 3 months)

Section 6: Comparative Analysis

markdown

## Comparative Analysis

### Performance vs Industry Standards

| System | HammerEngine | Industry Avg | Status |
|--------|--------------|--------------|--------|
| Entity Count @ 60 FPS | 10,000 | 5,000-8,000 | ✓ Above Avg |
| Collision Checks/sec | 125,000 | 80,000-100,000 | ✓ Above Avg |
| Event Throughput | 8,500/sec | 5,000-10,000 | ✓ Average |
| Memory/Entity | 13 KB | 10-20 KB | ✓ Average |

**Overall:** HammerEngine performs above industry averages for 2D game engines.

### Performance vs Project Goals

| Goal | Target | Current | Status |
|------|--------|---------|--------|
| 10K+ Entities @ 60 FPS | 60 FPS | 62.3 FPS | ✓ Exceeded |
| AI CPU Usage | <6% | 5.8% | ✓ Met |
| Event Throughput | 10K/sec | 8.5K/sec | ⚠ 85% of goal |
| Collision Efficiency | N/A | 94.2% hash | ✓ Excellent |
| Pathfinding Speed | <10ms | 8.5ms | ✓ Met |

**Overall Progress:** 80% of goals met or exceeded

Section 7: Technical Details

markdown

## Technical Details

### Test Environment

- **Hardware:** <CPU-model>, <RAM-size>
- **OS:** Linux 6.16.4 (Bazzite Fedora 42)
- **Compiler:** GCC/Clang <version>, C++20
- **Build Flags:** -O3 -flto -march=x86-64-v3 -mavx2
- **SDL Version:** SDL3 (latest)

### Benchmark Methodology

- **Duration:** 20 minutes total
- **Repetitions:** 5 runs per benchmark (median reported)
- **Warm-up:** 30 seconds per test
- **Isolation:** Tests run sequentially, system idle
- **Profiling:** Callgrind with 1% sampling

### Data Collection

- **Metrics Collection:** Automated via test scripts
- **Storage:** $PROJECT_ROOT/test_results/ directory
- **Baseline:** Updated monthly
- **History:** 6 months retained

### Reliability

- **FPS Variance:** 2.9% CoV (excellent)
- **CPU Variance:** 5.1% CoV (good)
- **Memory Variance:** 1.2% CoV (excellent)

**Overall:** Results are highly reliable and reproducible.

Section 8: Appendices

markdown

## Appendix A: Raw Metrics

### AI System Benchmark (Raw Data)

Entities: 10000 FPS: 62.3 CPU: 5.8% Update Time: 12.4ms Batch Processing: 2.1ms Behavior Updates: 8.3ms Memory Usage: 45 MB Thread Safety: Mutex-protected Double Buffer: Enabled Cache Efficiency: 92%

code


<Include raw data for all systems>

## Appendix B: Callgrind Full Output

<Full callgrind_annotate output if available> ```

Appendix C: Test Scripts

All benchmark tests are located in:

•tests/test_scripts/run_ai_system_benchmark.sh
•tests/test_scripts/run_collision_benchmark.sh
• <list all benchmark scripts>

Run full suite:

bash

./run_all_tests.sh --benchmarks-only

Appendix D: Baseline History

Date	AI FPS	Collision Checks	Event Throughput
2025-01-15	62.3	125,000	8,500
2024-12-15	61.8	120,000	8,200
2024-11-15	59.2	110,000	7,000

<Full historical baseline data> ```

Step 7: Format Report

Output Formats:

1. Markdown (Default)

Save to:

code

docs/performance_reports/performance_report_YYYY-MM-DD.md

Advantages:

•Easy to read in text editor
•Version control friendly
•Can be converted to other formats

2. HTML (Optional)

Convert markdown to HTML:

bash

# If pandoc available
pandoc performance_report.md -o performance_report.html \
  --standalone \
  --css=report_style.css \
  --metadata title="HammerEngine Performance Report"

Advantages:

•Interactive charts (if using JavaScript)
•Better for web viewing
•Professional appearance

3. PDF (Optional)

Convert markdown to PDF:

bash

# If pandoc + LaTeX available
pandoc performance_report.md -o performance_report.pdf \
  --pdf-engine=xelatex \
  --variable geometry:margin=1in \
  --variable fontsize=11pt \
  --metadata title="HammerEngine Performance Report"

Advantages:

•Professional document
•Easy to share with stakeholders
•Print-ready

Step 8: Generate Summary Output

Console Output:

code

=== HammerEngine Benchmark Report Generated ===

Report Date: 2025-01-15 14:30:22
Benchmarks Analyzed: 6 systems
Metrics Collected: 42 data points
Baseline Comparison: Available (2024-12-15)

Performance Status: ✓ EXCELLENT

Key Highlights:
  ✓ AI System: 10,000 entities @ 62.3 FPS
  ✓ Collision: 125,000 checks/sec
  ✓ Pathfinding: 8.5ms avg calculation
  ⚠ Event Dispatch: +8% latency (monitor)

Report Generated:
  📄 Markdown: docs/performance_reports/performance_report_2025-01-15.md
  🌐 HTML: docs/performance_reports/performance_report_2025-01-15.html
  📑 PDF: docs/performance_reports/performance_report_2025-01-15.pdf

Report Size: 2.3 MB (includes charts and raw data)
Generation Time: 2m 15s

Next Steps:
  1. Review optimization opportunities (Section 4)
  2. Address update thread budget overrun
  3. Monitor event dispatch latency trend
  4. Update baseline after validating improvements

---
Report ready for distribution

Usage Examples

When the user says:

•"generate performance report"
•"create benchmark report"
•"document current performance"
•"prepare performance analysis"
•"make stakeholder report"

Activate this Skill automatically.

Report Customization Options

Ask user for:

Performance Expectations

•Data Collection: 1-2 minutes
•Analysis: 30-60 seconds
•Report Generation: 30-60 seconds
•Total Time: 2-4 minutes

Manual equivalent: 45-60 minutes

Integration with Workflow

Use this Skill:

•After completing benchmark runs
•Before milestone releases
•For monthly performance tracking
•When preparing stakeholder updates
•To document optimization efforts
•Before/after major refactoring

File Management

Reports are saved to:

code

docs/performance_reports/
├── performance_report_2025-01-15.md
├── performance_report_2025-01-15.html
├── performance_report_2025-01-15.pdf
└── latest_report.md (symlink to most recent)

Cleanup old reports:

bash

# Keep last 12 months, archive older
find docs/performance_reports/ -name "*.md" -mtime +365 -exec mv {} archive/ \;

Quality Assurance

Report includes:

•✓ All benchmark systems analyzed
•✓ Statistical validity checks (min 3 data points)
•✓ Trend analysis (if historical data available)
•✓ Clear status indicators (✓/⚠/❌)
•✓ Actionable recommendations
•✓ Professional formatting
•✓ Version information (git commit, date)
•✓ Reproducibility information (test environment)

Advanced Features

ASCII Charts

Generate simple ASCII charts for trends:

code

FPS Trend:
68 ┤          ╭─
66 ┤        ╭─╯
64 ┤      ╭─╯
62 ┤    ╭─╯
60 ┤  ╭─╯
58 ┤╭─╯
   └─────────────

Sparklines

Compact trend indicators:

•AI FPS: ▁▂▃▅▆█ (improving)
•Collision: ███▇▅▃ (degrading)

Color Coding (in HTML/PDF)

•🟢 Green: Exceeds targets
•🟡 Yellow: Meets targets
•🟠 Orange: Below targets (warning)
•🔴 Red: Critical issues

Error Handling

If no benchmark data found:

•Report: "No benchmark data available. Run benchmarks first:"
•Show command: ./run_all_tests.sh --benchmarks-only

If baseline missing:

•Note in report: "No baseline for comparison. This will serve as baseline."
•Save current metrics as baseline

If incomplete data:

•Note missing systems in report
•Generate partial report
•Recommend running missing benchmarks

Exit Codes

•0: Report generated successfully
•1: No benchmark data found
•2: Report generation failed
•3: Partial report (missing data)

Time Savings

Manual Process: ~45-60 minutes

•Collect metrics: 10 min
•Calculate statistics: 15 min
•Format tables: 10 min
•Write analysis: 15 min
•Generate charts: 10 min

With This Skill: ~2-4 minutes

•Automated collection
•Automated analysis
•Automated formatting

Total Time Saved: ~40-55 minutes per report