[README.md]
Visual Summary
Left panel: Quantifies “market position” by combining vector depth, openness, and activity. → UA-RNG v1.7 sits at the top with the only full (3/3) composite rating.
Right panel: Radar chart compares structural capability profiles. → UA-RNG’s polygon covers every axis — the most balanced and complete project in the field. → Intel SVRNG spikes hard on vector breadth but collapses on openness and transparency.
A high-precision visualization and analytics toolkit for multi-compiler, multi-OS RNG performance benchmarking.
This project provides a clean, extensible, and publication-grade benchmarking dashboard for analyzing SIMD-optimized RNG performance across compilers and operating systems.
It consolidates CSV results, calculates throughput and variance deltas, and exports clear visual and tabular reports for cross-comparison.
- 🧮 2×2 Multi-Panel Dashboard — U64/F64 throughput, Speedup %, Δ-Variance
- 📊 Automatic Baseline Comparison — relative % improvement over a reference compiler
- 🧾 Aligned Markdown Reports — perfect ASCII tables for GitHub or terminal use
- 📈 Top-3 LaTeX Export — ready for paper or report inclusion
- 🧠 No LaTeX dependency — pure Matplotlib PDF backend
- ⚙️ Consistent Cross-Platform Results — stable across Linux, Windows, and MSYS2 builds
python -m venv venv
source venv/bin/activate
pip install -r requirements.txtUsage
python plot-cross-platform-adv.py \
--baseline windows-msvc \
--multi-panel \
--top3-table \
--style paper \
--output results/cross_compiler_full_report.pdf
Optional Flags
Flag Description
--baseline Reference platform for relative % speedups
--multi-panel Enables 2×2 composite plot layout
--top3-table Exports results/top3_table.tex
--style Choose paper (default) or dark theme
--output Specify output file path
Directory Structure
simd-bench-visualizer/
├── plot-cross-platform-adv.py # Main script
├── requirements.txt # Dependencies
├── results/
│ ├── bench_all_linux-gcc.csv
│ ├── bench_all_windows-msvc.csv
│ ├── cross_compiler_full_report.pdf
│ ├── cross_compiler_full_report_all_comparisons_pretty.md
│ └── top3_table.tex
├── CHANGELOG.md
├── CONTRIBUTING.md
├── CODE_OF_CONDUCT.md
├── SECURITY.md
├── LICENSE
└── README.md
Example Output
Multi-Panel Dashboard
U64 & F64 throughput per generator
Speedup vs baseline
Δ Variance (%)
ASCII Table Example
+
+
+
+
+
-+
| Generator | linux-gcc Read (Mops/s) | windows-msvc Read (Mops/s) | Winner | Advantage (%) |
+
+
+
+
+
-+
| std_mt19937 | 1014458959.86 | 495258380.79 | linux-gcc | 104.83 |
| std_mt19937_64 | 1200113371.71 | 692907957.14 | linux-gcc | 73.20 |
| minstd_rand | 1182112151.14 | 1003158884.63 | linux-gcc | 17.84 |
+
+
+
+
+
-+
Requirements
See requirements.txt :
numpy>=1.26.0,<2.0.0 pandas>=2.2.0,<3.0.0 matplotlib>=3.8.0,<4.0.0 seaborn>=0.13.0,<0.14.0 cycler>=0.12.0 fonttools>=4.40.0 kiwisolver>=1.4.5 pyparsing>=3.1.0 Pillow>=10.0.0 packaging>=23.0
Contributing
Pull requests and feature suggestions are welcome! See CONTRIBUTING.md for coding style and workflow.
License
MIT License — (c) 2025 RYOModular / whisprer See LICENSE for full text.
plaintext ref.
whisprer, W. (2025). Cross-Platform SIMD RNG Benchmark Visualizer (Version 1.0.0) [Computer software]. whisper.dev. https://github.com/whisprer/benchmark
Built with precision and pragmatism by woflfren “Benchmark everything. Optimize what matters.”