Line Plots

Complete reference for line plot visualizations in RING-5.

Overview

Line plots show trends and relationships over continuous variables. Essential for time-series analysis and tracking changes.

Configuration

Required:

• X-axis: Continuous variable (time, iteration, address)
• Y-axis: Metric (ipc, throughput, latency)

Optional:

• Trace by: Multiple series (config, benchmark)
• Line style: solid, dashed, dotted
• Markers: points, circles, squares

Use Cases

Time-Series Analysis

Track performance over simulation time:

X: simTicks
Y: ipc
Trace by: config
# Shows IPC evolution over time

Convergence Plots

Monitor algorithm convergence:

X: iteration
Y: error_rate
Trace by: algorithm
# Shows convergence speed

Workload Progression

Analyze behavior across workload phases:

X: instruction_count
Y: cache_miss_rate
Trace by: cache_level
# Shows miss rate progression

Styling

Line Properties

• Width: 2-3px for primary lines
• Style: Solid for main, dashed for reference
• Opacity: 0.8-1.0

Markers

• Show for sparse data (< 50 points)
• Hide for dense data
• Size: 6-8px

Colors

Use distinct colors per trace:

• Maximum 7 traces per plot
• Colorblind-friendly palette
• Consistent across related plots

Data Preparation

Filtering

Remove warmup phase:

{
    "type": "conditionSelector",
    "column": "phase",
    "mode": "not_equals",
    "threshold": "warmup"
}

Sorting

Ensure chronological order:

{
    "type": "sort",
    "order_dict": {"simTicks": "ascending"}
}

Best Practices

1. Limit to 5-7 traces per plot
2. Use consistent time units
3. Add reference lines for baselines
4. Label key events/phases
5. Use log scale for exponential data

Common Patterns

Multi-Config Comparison

X: time
Y: throughput
Trace by: config
Pipeline: Filter → Sort

Phase Analysis

X: phase_id
Y: ipc
Trace by: benchmark
Pipeline: Aggregate by phase

Next Steps

• Scatter Plots: Scatter-Plots.md
• Bar Charts: Bar-Charts.md