What is bench.sh?

What is bench.sh?

Bench.sh is a lightweight shell script designed to quickly benchmark the performance of a VPS or dedicated server. With just a single command, you can get an initial overview of the server's CPU capabilities, disk read/write speed (I/O), and network connection (download/upload) speed.

Definition and Function

Bench.sh collects data by running a series of simple tests: calculating prime numbers for CPU, writing and reading temporary files for I/O, and fetching files from several servers in various locations for network speed. The results are displayed directly in the terminal in an easy-to-read table format.

Its main purpose is for quick diagnostics before or after purchasing a VPS, comparing hosting providers, or ensuring the server is not in poor condition due to noisy neighbor issues.

Advantages of Using bench.sh

• Simple and portable: Only needs an internet connection and a standard shell. No additional package installation required.
• Fast: The entire process usually completes in 1–3 minutes.
• Free and open source: The code can be inspected for security.
• Support for multiple network test locations: Displays speed to servers in the US, Europe, Asia, etc.
• Easy to share: Results can be directly copied and pasted to forums or chats.

---

How to Download and Run bench.sh

The process is very simple. You just need to access the server terminal (SSH) and run a curl or wget command pointing to the official script.

Basic Terminal Command

Run the following command:

curl -sL yabs.sh | bash

Or if you prefer wget:

wget -qO- yabs.sh | bash

The script will automatically download and execute the benchmark. Wait until all tests are complete — the results will appear on screen.

Useful Additional Options

Bench.sh (or variants like yabs.sh) supports several parameters:

• --skip-io : Skips the disk I/O test (useful if you only want to check CPU and network).
• --skip-network : Skips the network test.
• --skip-cpu : Skips the CPU test.
• --location : Specifies the location of the network server to test (example: --location US).

Example usage:

curl -sL yabs.sh | bash -s -- --skip-io

With these options, you can shorten the time or focus on specific aspects.

---

Understanding the CPU, I/O, and Network Benchmark Results

Once the script finishes, you will see three main sections: CPU, Disk I/O, and Network Speed. Each has numbers that you need to know how to read.

Interpreting CPU and I/O Scores

The CPU test uses a method of calculating prime numbers in GCD (or a similar algorithm). The score is usually displayed in events per second (or similar). The higher the number, the faster the processor.

• For entry-level VPS (1 core ~2.0 GHz), CPU scores typically range between 500–2000 events/s.
• Mid-range VPS (2–4 modern cores) can reach 4000–8000 events/s.
• High-end dedicated servers can be above 15,000 events/s.

The Disk I/O test writes and reads files of a certain size (usually 4K and 1M) and displays speed in MB/s. Pay attention to two values:

• 4K block: measures random access speed. VPS with SSD should be > 20 MB/s. HDD is usually < 5 MB/s.
• 1M block: measures sequential speed. SSD can reach > 500 MB/s, HDD in the range of 100–200 MB/s.

If the I/O is very low, the provider may be using shared storage that is overloaded or using old HDDs.

Network Speed Analysis (Download/Upload)

The network section displays a list of servers from various countries (e.g., Amsterdam, Singapore, New York) with download and upload speeds in Mbps.

• Watch for consistency: If one location is very slow, that may be normal due to geographical distance or poor peering. But if all locations are slow (below 100 Mbps for a VPS with 1 Gbps bandwidth), there is a problem.
• Upload is usually lower than download due to upstream provider limitations. Compare with the promised specifications (e.g., dedicated 1 Gbps). If the result is only 300 Mbps under idle load, suspect throttling.

---

Tips and Recommendations for Benchmark Results

Raw numbers mean nothing without context. Here's how to use them properly.

Comparing with Other VPS

You can compare results with public databases (e.g., on LowEndTalk forums, VPSBenchmarks, or sites like HostScore). Pay attention to:

• Identical specifications: Compare with VPS that have the same number of cores, RAM, and storage type.
• Testing time: Run benchmarks during peak and off-peak hours to see fluctuations.
• Location: Servers in Europe usually yield different results than servers in Asia due to different network routes.

Use the CPU and I/O scores to assess whether the VPS is worth the price paid.

When to Rerun the Benchmark

Don't rely on a single test. Repeat the benchmark in the following situations:

• After a server restart or maintenance to ensure no changes.
• When the server feels slow — compare the new results with previous ones.
• After migration or upgrade to measure improvement.
• After a few months — performance can degrade over time due to node saturation or hardware degradation.

Run at least twice at different times (e.g., at night and during the day) to assess stability.

---

Bench.sh is not a tool for in-depth analysis, but it is reliable enough as a quick health check. Use its results as an initial indicator before concluding server issues — and always compare with the context of your specifications.

This article was written by artificial intelligence (AI) using the dahono/deepseek-v4-flash model via Dahono Labs.

This article was translated by Artificial Intelligence (AI) using dahono/deepseek-v4-flash via Dahono Labs.