Creating Plots Guide

Comprehensive guide to creating and customizing visualizations in RING-5.

Overview

RING-5 provides seven plot types for gem5 data visualization:

• Bar Chart
• Grouped Bar Chart
• Stacked Bar Chart
• Grouped Stacked Bar Chart
• Line Plot
• Scatter Plot
• Histogram

All plots are built with Plotly for interactivity and publication quality.

Getting Started

Prerequisites

• Data loaded via Data Source or Upload Data
• Navigate to Manage Plots page

Creating Your First Plot

1. Click Create New Plot
2. Enter plot name (e.g., “IPC Comparison”)
3. Select plot type from dropdown
4. Click Create

Plot Types

Bar Chart

Purpose: Compare values across categories

Configuration:

• X-axis: Categorical variable (e.g., config, benchmark)
• Y-axis: Numeric variable (e.g., ipc, throughput)

Use Cases:

• Compare configurations
• Show single metric per category
• Simple performance comparisons

Grouped Bar Chart

Purpose: Compare multiple groups side-by-side

Configuration:

• X-axis: Primary category (e.g., benchmark)
• Y-axis: Numeric variable (e.g., ipc)
• Group by: Secondary category (e.g., config)

Use Cases:

• Multi-configuration comparisons per benchmark
• Side-by-side performance metrics
• A/B testing results

Stacked Bar Chart

Purpose: Show composition of totals

Configuration:

• X-axis: Category (e.g., benchmark)
• Y-axis: Numeric variable (e.g., cache_misses)
• Stack by: Component (e.g., cache_level)

Use Cases:

• Cache hierarchy breakdown
• Memory allocation visualization
• Component contribution analysis

Grouped Stacked Bar Chart

Purpose: Combine grouping and stacking

Configuration:

• X-axis: Primary category
• Y-axis: Numeric variable
• Group by: Groups (e.g., config)
• Stack by: Stack components (e.g., operation_type)

Use Cases:

• Complex multi-dimensional comparisons
• Showing both configuration and breakdown
• Advanced performance analysis

Line Plot

Purpose: Show trends over continuous variables

Configuration:

• X-axis: Continuous variable (e.g., time, iteration)
• Y-axis: Metric (e.g., ipc, throughput)
• Trace by: Multiple series (e.g., config)

Use Cases:

• Time-series analysis
• Convergence plots
• Performance over workload progression

Scatter Plot

Purpose: Visualize relationships between two variables

Configuration:

• X-axis: First variable (e.g., cache_miss_rate)
• Y-axis: Second variable (e.g., ipc)
• Color by: Category (e.g., config)
• Size by: Third variable (optional)

Use Cases:

• Correlation analysis
• Identifying outliers
• Multi-dimensional visualization

Histogram

Purpose: Show distribution of values

Configuration:

• Values: Variable to analyze (e.g., ipc)
• Bins: Number of bins (default: auto)
• Group by: Optional grouping (e.g., config)

Use Cases:

• Distribution analysis
• Identifying data patterns
• Performance variability assessment

Plot Configuration

Data Mapping

Map DataFrame columns to plot axes:

1. Select Columns: Use dropdowns to select columns
2. Preview: Check column contents
3. Validate: Ensure correct data types

Tips:

• X-axis: Usually categorical or time-series
• Y-axis: Always numeric
• Group/Stack: Categorical variables

Styling Options

Customize plot appearance with title, labels, legend, colors, size, and fonts.

Layout Options

Control plot layout with margins, grid lines, and axis properties.

Data Processing Pipeline

Each plot has an independent pipeline for transformations. Order matters: Column Selector → Filter → Sort → Normalize → Mean.

Best Practices

1. Choose the right plot type: Bar for comparison, Stacked for composition, Line for trends
2. Clean Data First: Use Data Managers
3. Publication Quality: Use 14pt+ fonts
4. Clear Labels: Descriptive axis titles

Next Steps

• Plot Type References: Detailed guides for each plot type in plots/
• Data Transformations: Master Shapers
• API Reference: See Plotting API

Need Help? Check Troubleshooting or open a GitHub issue.