TreeScape Documentation

TreeScape is a Jupyter-based visualization tool for performance data, enabling users to programmatically render graphs. With TreeScape, you can load an ensemble of Caliper performance files and visualize the collective performance of an application across many runs.

Features

Multiple Data Readers: Support for both CaliperReader and ThicketReader
Interactive Visualizations: Stacked line charts with integrated flame graphs
Static Plotting: Matplotlib-based multi-line charts for publication
Performance Tracking: Track performance changes over time across multiple runs
Flexible Data Model: Filter, sort, and analyze performance data programmatically
Export Capabilities: Export visualizations to SVG format

Quick Start

Installation

git clone https://github.com/LLNL/treescape.git
cd treescape
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Basic Usage

import treescape as ts

# Load Caliper performance files
reader = ts.CaliReader("/path/to/cali/files")

# Create a data model
model = ts.TreeScapeModel(reader)

# Create an interactive visualization
viz = ts.StackedLine()
viz.setXAxis("launchdate")
viz.setYAxis("avg")
viz.setDrillLevel(["main", "LagrangeLeapFrog"])
viz.render(model)

# Or create a static matplotlib plot
ml = ts.MultiLine(model)
ml.plot_sums("launchdate", "main", "test")

Table of Contents

User Guide

Additional Information

Indices and tables