How do I profile Lua code in NeoVim with PerfAnno?

Use the built-in LuaJIT profiler commands :PerfLuaProfileStart and :PerfLuaProfileStop to capture stack traces directly within NeoVim, then load and annotate the data without external tools, as detailed in the Lua profiling section.

Can PerfAnno work with Python profilers like cProfile?

Not directly; PerfAnno requires output in perf or flamegraph format, so you'd need to convert cProfile data or use a compatible profiler. It's language-agnostic but depends on supported profiler outputs.

Is PerfAnno better than using perf report in the terminal?

PerfAnno integrates profiling into NeoVim with visual annotations and interactive navigation, offering a more seamless workflow for developers already in the editor, while perf report is a separate TUI tool with different visualization strengths.

How to set up keybindings for PerfAnno quickly?

Copy the example config from the README to map leader keys to commands like :PerfAnnotate and :PerfHottestLines, ensuring easy access to annotation and navigation features without memorizing commands.

What's needed to analyze multi-threaded apps with PerfAnno?

Enable thread_support in setup and load perf data recorded with --call-graph dwarf; you'll be prompted to select threads for analysis, allowing per-thread or aggregated performance insights.

How does PerfAnno handle large profiling data files?

Loading call graph data can be slow, but the experimental caching feature lets you save and restore profiles to speed up repeated analysis, though it's not fully robust as noted in the future goals.

perfanno.nvim

MITLua

A NeoVim plugin that annotates source code with profiling data from perf, LuaJIT, and other profilers to visualize performance bottlenecks.

GitHub

What is perfanno.nvim?

PerfAnno is a NeoVim plugin written in Lua that integrates profiling data directly into the editor. It reads output from profilers like perf and LuaJIT, then annotates source code lines with performance metrics such as CPU cycles, cache misses, and branch mispredictions, allowing developers to visually identify hotspots and explore call graphs without leaving their coding environment.

Target Audience

NeoVim users who are performance-tuning applications in languages like C, C++, Lua, or Python, particularly those working on multi-threaded applications or needing detailed call graph analysis.

Value Proposition

Developers choose PerfAnno because it brings profiler data directly into NeoVim, enabling real-time, context-aware performance analysis with features like interactive call graph navigation, multi-profiler support, and customizable annotations, eliminating the need to switch between tools.

Overview

NeoVim lua plugin that annotates source code with profiling information from perf, LuaJIT, or other profilers.

Use Cases

Best For

Performance profiling C/C++ applications with perf call-graph data directly in NeoVim.

Related Projects

Community-curated · Updated weekly · 100% open source

Found a gem we're missing?

Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.

Submit a project Star on GitHub

365 stars6 forks0 contributors

Profiling Lua code within NeoVim itself using the built-in LuaJIT profiler integration.

Analyzing multi-threaded application performance with per-thread profiling data.

Identifying performance hotspots via visual code annotations with metrics like CPU cycles or cache misses.

Exploring call graphs interactively using Telescope or fzf-lua to navigate between callers and callees.

Profiling applications with custom profilers that output flamegraph format stack traces.

Not Ideal For

Developers using traditional Vim or editors other than NeoVim, as the plugin is specifically built for NeoVim 0.10+.
Teams needing GUI-based profiling visualizations like interactive flame graphs or heat maps beyond text annotations in an editor.
Projects relying on profilers that don't output perf data or flamegraph format, such as Valgrind or proprietary tools without conversion steps.

Pros & Cons

Pros

In-Editor Profiling Integration

Directly annotates source code lines in NeoVim with performance metrics like CPU cycles, eliminating the need to switch between profiler and editor for real-time analysis.

Flexible Profiler Support

Works with multiple profilers including perf (call-graph and flat modes), LuaJIT, and any tool outputting flamegraph format, making it versatile for C, C++, Lua, and Python projects.

Interactive Call Graph Navigation

Integrates with Telescope or fzf-lua to jump to hottest lines and explore callers/callees interactively, with hotkeys for dynamic navigation within finders.

Customizable Visual Feedback

Allows configuration of line highlights and annotation formats via Lua, enabling users to tailor displays to their color schemes and preferences, as shown in the example config.

Cons

Complex Initial Setup

Requires manual configuration of highlights, keybindings, and dependencies like Telescope or fzf-lua for full functionality, adding overhead for new users beyond simple plugin installation.

Incomplete and Experimental Features

Features such as caching are labeled experimental, and commands like :PerfCycleFormat have known issues (e.g., resetting relative annotations), reducing reliability for some workflows.

Perf Version and Data Dependency

Call graph mode may require recent perf versions, and loading large perf data can be slow; the plugin also depends on accurate debug symbols for precise line annotations, which can be a hurdle.

Frequently Asked Questions

Home

Neovim

nvim-dap

Debug Adapter Protocol client implementation for Neovim

Stars7,110

Forks260

Last commit1 month ago

nvim-dap-ui

A UI for nvim-dap

Stars3,332

Forks125

Last commit1 month ago

nvim-dap-virtual-text

nvim-dap-virtual-text enhances the debugging experience in Neovim by showing variable values inline as virtual text while using nvim-dap. It integrates with nvim-treesitter to intelligently locate variable definitions and references, making debugging more visual and efficient. ## Key Features - **Inline Variable Display** — Shows variable values directly in the code as virtual text during debugging. - **Changed Value Highlighting** — Highlights variables that have changed since the last step with customizable colors. - **Exception Stop Reasons** — Displays the reason for debugger stops due to exceptions. - **Multi-language Support** — Works with any language supported by nvim-treesitter's locals.scm queries (C/C++, Python, Rust, Go, Java, etc.). - **Customizable Display** — Offers extensive configuration options for virtual text position, formatting, and behavior. - **Frame-aware Display** — Can show variable values from all stack frames, not just the current one. ## Philosophy The plugin follows a minimalistic approach, enhancing the debugging workflow without cluttering the interface, and leverages existing Neovim ecosystems like nvim-treesitter for intelligent parsing.

Stars1,071

Forks35

Last commit11 months ago