trueno-viz

Table of Contents

• Installation
• Usage
• Architecture
• API Reference
• Examples
• Testing
• Contributing
• License

Features

• 5.6x SIMD Speedup - Real platform intrinsics (SSE2/AVX2/NEON)
• 8ms Frame Time - 2X faster than btop's 16ms target
• GPU Compute - CUDA, Vulkan, Metal, and WebGPU
• Zero Dependencies - Pure Rust rendering to PNG, SVG, or terminal

ttop - Terminal Top

The flagship application. A pure Rust system monitor that's 10X better than btop.

cargo install ttop

Why ttop?

Panels

Advanced Features

PSI Pressure:    CPU ○ 2.1%  MEM ○ 0.0%  I/O ◔ 1.5%
Containers:      ▶  5.2%  512M web-app
Connections:     ESTAB TCP nginx → 192.168.1.100:443
File Treemap:    Pareto-style large file visualization

Platform Support

Full ttop documentation →

trueno-viz Library

Installation

[dependencies]
trueno-viz = "0.2"

# With GPU acceleration
trueno-viz = { version = "0.2", features = ["gpu"] }

# With system monitoring (powers ttop)
trueno-viz = { version = "0.2", features = ["monitor"] }

Quick Start

use trueno_viz::prelude::*;
use trueno_viz::output::{TerminalEncoder, TerminalMode};

let plot = ScatterPlot::new()
    .x(&[1.0, 2.0, 3.0, 4.0])
    .y(&[1.0, 4.0, 2.0, 8.0])
    .build()?;

let fb = plot.to_framebuffer()?;
TerminalEncoder::new()
    .mode(TerminalMode::Ascii)
    .print(&fb);

Plot Types

• Scatter - Point clouds with regression lines
• Line - Time series and trends
• Histogram - Distribution analysis
• Heatmap - Correlation matrices
• Box/Violin - Statistical distributions
• ROC/PR - ML model evaluation
• Confusion Matrix - Classification results
• Loss Curves - Training visualization

Output Formats

SIMD Collectors

Real platform intrinsics for system monitoring (v0.1.14+):

┌─────────────────────────────────────────────────────────────┐
│  SSE2/AVX2 (x86_64)     │  std::arch::x86_64 intrinsics    │
│  NEON (aarch64)         │  std::arch::aarch64 intrinsics   │
│  5.6x measured speedup  │  Byte scanning operations        │
└─────────────────────────────────────────────────────────────┘

Three-Tier Storage

Hot   → SimdRingBuffer  (O(1) statistics)
Warm  → Compressed      (LZ4 in-memory)
Cold  → Disk            (Persistent history)

Performance

Examples

# ttop - system monitor
cargo install ttop && ttop

# Visualization examples
cargo run --example scatter_basic
cargo run --example heatmap_correlation
cargo run --example loss_training
cargo run --example confusion_matrix_ml
cargo run --example roc_pr_curves
cargo run --example terminal_output

# All examples
cargo run --example readme_demo
cargo run --example grammar_of_graphics

Architecture

trueno-viz/
├── src/
│   ├── framebuffer.rs      # SIMD-aligned pixel buffer (64-byte AVX-512)
│   ├── plots/              # Scatter, Line, Heatmap, Histogram, etc.
│   ├── grammar/            # Grammar of Graphics (ggplot-style API)
│   ├── output/             # PNG, SVG, Terminal encoders
│   └── monitor/            # System monitoring (feature: monitor)
│       └── simd/           # Real SIMD kernels
│           ├── kernels.rs      # Platform intrinsics
│           ├── ring_buffer.rs  # O(1) statistics
│           ├── timeseries.rs   # Three-tier storage
│           └── correlation.rs  # Pearson correlation
└── crates/
    └── ttop/               # System monitor binary (10X better than btop)

Testing

cargo test --lib          # Unit tests
cargo test                # All tests including integration
cargo test -p ttop        # ttop system monitor tests
make coverage             # Coverage report
make bench                # SIMD performance benchmarks

Property-based tests cover framebuffer operations, SIMD kernel correctness, and plot rendering invariants.

Documentation

• API Documentation - Complete API reference
• The Visualization Guide - mdBook with tutorials
• ttop README - System monitor documentation
• Examples - Runnable code samples

Contributing

Contributions are welcome! Please see the CONTRIBUTING.md guide for details.

MSRV

Minimum Supported Rust Version: 1.75

See Also

• Cookbook — 18 runnable examples

License

MIT OR Apache-2.0

Part of the Aprender monorepo — 70 workspace crates.