🧬 Calyx: Simple & Precise Linux System Benchmarking

Calyx is a lightweight, all-in-one system benchmarking and diagnostic tool for Linux. It is designed as a high-performance, fully native C++23 alternative to classic scripts like bench.sh or yabs.sh.

If you've ever wondered "What are my server's specs?", "How much RAM is currently used?", "How fast is my disk?", or "What is my actual internet speed?", Calyx gives you the answers instantly. It runs directly in your terminal and produces a beautiful, easy-to-read hardware and network performance report.

---

🌟 What makes Calyx different?

Many benchmarking tools are either too complicated for beginners or too inaccurate because of system cache "cheats". Calyx solves both:

• Beginner-Friendly: No installation, no complex arguments. Just run one command, and it shows you everything you need to know.
• Bypasses Cache "Cheating": When testing disk speeds, Linux normally caches files in RAM, making slow disks look artificially fast. Calyx uses modern Linux io_uring and O_DIRECT to bypass this, measuring your real hardware capability.
• Intelligent & Adaptive: Calyx automatically tunes itself to your Linux kernel. It probes your system limits and configures the fastest possible path automatically.
• Ultra-Portable: It compiles down to a single tiny file (~3 MB) that has no dependencies. You can copy it to any Linux machine and run it instantly.

---

📊 What does it measure?

When you run Calyx, it tests four main areas of your system:

1. 🧠 CPU & Hardware

• Detects your CPU model name, active cores, and max clock frequency.
• Checks if security hardware acceleration (AES-NI) is enabled.
• Checks if Hardware Virtualization is supported.

2. 🐧 OS & System Information

• Identifies your Linux distribution, kernel version, and virtualization type (e.g., Docker, KVM, Hyper-V).
• Monitors uptime and system load averages.
• Displays your current TCP Congestion Control algorithm (e.g., bbr, cubic).

3. 💾 Storage, Memory & ZSwap

• Lists partition sizes, used space, and filesystem type.
• Measures real RAM and Swap usage.
• Disk I/O Performance: Measures sequential read and write speeds using Linux-native io_uring and O_DIRECT to ensure cache-bypass accuracy.
• ZSwap Metrics: Shows compression ratio, and helps you identify memory pressure:
  • Spilled: Data pushed to disk because compressed memory was full (high values mean you need more RAM!).
  • Rejected: Data that failed to compress because the CPU was too busy.
  • Capped: Data rejected because the ZSwap pool size limit was hit.

4. 🌐 Internet & Network Speed

• Performs an instant online check and ISP auto-lookup.
• Tests download speed, upload speed, latency (ping), and packet loss to multiple regional and international nodes.

---

🚀 Quick Start (No Install Required)

You don't need to install or compile anything to use Calyx. Just copy and paste this command into your Linux terminal:

bash <(curl -fsL https://calyx.pages.dev/run)

Note

This command automatically detects your system architecture, downloads the secure static binary, runs the benchmark, displays your report, and cleans up afterwards leaving your system clean.

💡 Troubleshooting: "io_uring fixed buffers disabled"

If you see Performance Hint: io_uring fixed buffers disabled (Memory limit), using fallback. in your storage test:

This is a very common hint on VPS servers or non-root accounts. It means Calyx wanted to use locked memory buffers for zero-copy high-performance I/O, but your system's locked memory limit (ulimit -l) was too low.

How to get maximum performance:

• Run as root (using sudo or as root user), which automatically bypasses the memory lock limits.
• Increase the limit manually by running ulimit -l unlimited in your terminal session before running the benchmark.

---

🛠️ Building from Source

If you prefer to compile Calyx yourself, you can build a fully static, portable binary using Docker.

Requirements

• Operating System: Linux (any distribution)
• Architecture: x86_64 (Intel/AMD) and aarch64 (ARM64)
• Kernel: Version 5.10 or newer (required for modern disk benchmarks)
• Docker: Version 20.10 or newer (Version 23.0 or newer recommended)

Build Steps

1. Clone the repository:

   git clone https://github.com/relvinarsenio/calyx.git
   cd calyx

2. Run the static build script:

   chmod +x build-static.sh
   ./build-static.sh

Once completed, your single portable binary will be generated at ./dist/calyx. You can run it locally:

./dist/calyx

---

📊 Example Output

Here is what the terminal output looks like when Calyx finishes running:

──────────────────────── Calyx - Linux System Benchmarking Utility (v1.1.0) ────────────────────────
 Author             : Alfie Ardinata (https://calyx.pages.dev/)
 GitHub             : https://github.com/relvinarsenio/calyx
 Usage              : ./calyx
────────────────────────────────────────────────────────────────────────────────────────────────────
  -> CPU & Hardware
  CPU Model            : AMD Ryzen 5 7535HS with Radeon Graphics
  CPU Cores            : 6 @ 4584.2 MHz (Max)
  CPU Cache            : 16 MB
  AES-NI               : ✓ Enabled
  Hardware Virt        : ✗ Disabled
 
  -> System Info
  OS                   : Debian GNU/Linux 13 (trixie)
  Arch                 : x86_64 (64 Bit)
  Kernel               : 6.19.6+deb14-amd64
  TCP CC               : bbr
  Virtualization       : Hyper-V
  System Uptime        : 14 hours, 38 mins
  Load Average         : 1.54, 1.06, 0.97
 
  -> Storage & Memory
  Test Path            : /home/user/Github/calyx (/dev/sda2 (xfs))
  Size Partition       : 63 GB (25 GB Used)
  Total Mem            : 6.9 GB (4.5 GB Used)
  Total Swap           : 5.2 GB (3.1 GB Used)
    -> Partition        : 4 GB (2.8 GB Used) (/dev/sdb)
    -> ZSwap            : 1.2 GB → 308.1 MB (3.88×) [zstd, limit: 1.4 GB (20%)]
                          Spilled: 10.8 GB  Rejected: 22.1 GB  Capped: 157.1 MB
 
  -> Network
  IPv4/IPv6            : ✓ Online / ✗ Offline
  ISP                  : AS13335 Cloudflare, Inc.
  Location             : Bandar Lampung / ID
  Region               : Lampung
────────────────────────────────────────────────────────────────────────────────────────────────────
Running I/O Test (1 GB File, Seq 1M Q16T1)...
 [ Throughput ]
   I/O Speed (Run #1)  :  Write    2.96 GB/s  │  Read    4.47 GB/s
   I/O Speed (Run #2)  :  Write    3.13 GB/s  │  Read    4.41 GB/s
   I/O Speed (Run #3)  :  Write    3.21 GB/s  │  Read    4.47 GB/s
   I/O Speed (Average) :  Write    3.11 GB/s  │  Read    4.45 GB/s

 [ Write Latency ]
   • Summary    :  Avg:    5.02 ms │ Min:    1.93 ms │ Max:    16.7 ms
   • Percentile :  p50:     4.2 ms │ p95:    9.51 ms │ p99:    10.6 ms │ p99.9:   16.59 ms

 [ Read Latency ]
   • Summary    :  Avg:    3.48 ms │ Min:    1.43 ms │ Max:    5.44 ms
   • Percentile :  p50:    3.49 ms │ p95:    3.96 ms │ p99:    4.47 ms │ p99.9:    5.25 ms
────────────────────────────────────────────────────────────────────────────────────────────────────
 Downloading Speedtest CLI...
  Node Name              Download          Upload            Latency     Loss    
  Speedtest.net (Auto)   138.91 Mbps       44.68 Mbps        19.87 ms    0.00 %  
  Singapore, SG          157.43 Mbps       43.07 Mbps        30.02 ms    0.00 %  
  Los Angeles, US        154.83 Mbps       16.60 Mbps        229.35 ms   0.00 %  
  Montreal, CA           145.71 Mbps       15.15 Mbps        267.46 ms   0.00 %  
  London, UK             147.87 Mbps       22.79 Mbps        212.89 ms   0.00 %  
  Amsterdam, NL          131.58 Mbps       16.16 Mbps        282.82 ms   0.00 %  
  Sydney, AU             144.35 Mbps       15.17 Mbps        235.15 ms   0.00 %  
 ────────────────────────────────────────────────────────────────────────────────────────────────────
  Finished in        : 3 min 47 sec

---

🤝 Contributing & Support

We welcome contributions! Whether it's adding new features, fixing bugs, or improving documentation, check out our Contributing Guidelines to get started.

Found a bug or have a feature request? Please open an issue on our GitHub repository.

---

⚖️ License & Code of Conduct

• This project is licensed under the Mozilla Public License 2.0.
• Please note that this project is released with a Code of Conduct. By participating, you agree to abide by its terms.