HPCtoolkit DataFrame

Operate on HPCtoolkit XML database files as pandas DataFrames.

Database files generated by HPCtoolkit can be read by the GUI-based tools provided by developers of HPCtoolkit. However, programmatic access and analysis of such files is troublesome.

This library provides an HPCtoolkitDataFrame object, which is essentially a pandas DataFrame and can be queried and sliced as easily as any DataFrame. But it extends this functionality with methods for analysis and visualisation of performance data.

Contents

• Usage
• Installation
  • Requirements
• Links

Usage

Please see examples.ipynb for details.

Installation

For simplest installation use pip:

pip3 install hpctoolkit_dataframe

Requirements

Python version 3.8 or later.

Python libraries as specified in requirements.txt.

Building and running tests additionally requires packages listed in requirements_test.txt.

Tested on Linux, macOS and Windows.

Links

• HPCtoolkit: http://hpctoolkit.org/
• pandas.DataFrame: https://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.html