LANBench is a minimalist, open-source network benchmarking utility. Unlike popular tools like iPerf3 or Netperf—which rely heavily on the operating system's TCP/IP stack and can be limited by CPU performance at high speeds—LANBench is designed to test the raw maximum achievable throughput of a network link by minimizing software overhead.
It achieves this by using a simple, custom protocol over raw sockets (or high-performance transports) and focusing on pure data pushing/pulling without complex application-layer logic. The primary goal is to answer the question: "What is the absolute maximum bandwidth this network cable, switch, and NIC combination can provide, ignoring OS protocol stack bottlenecks?"
LANBench provides a systematic way to quantify LAN performance, helping engineers make data-driven decisions about hardware, configuration, and capacity. When used with careful test planning and attention to test endpoint limitations, it delivers actionable insights into throughput, latency, and reliability of local networks.
Would you like a shorter summary, a sample CLI test script, or a one-page printable report template?
LANBench is a free, portable network benchmarking utility designed to test the performance of a local area network (LAN). It specifically measures TCP performance between two computers using the Winsock 2.2 API. Key Features
Customizable Tests: You can specify packet sizes, test duration, and connection counts. LANBench
Flexible Modes: Supports "Send Only," "Receive Only," or simultaneous "Send and Receive" (full-duplex) testing.
Portable Utility: Does not require installation, making it easy to run from a USB drive. How to Use It
To measure network speed, you must run the tool on two separate computers simultaneously:
Server Mode: Set one computer to "Listen" on a specific port (default is 8988).
Client Mode: Set the second computer to "Client" and enter the IP address of the server computer. Standard benchmarks would report "100 t/s
Run Benchmark: Start the test from the client to see real-time throughput data. Why Use LANBench?
Verify Hardware: Use it to see if your Gigabit network is actually hitting its ~125 MB/s theoretical limit.
Identify Bottlenecks: Compare performance between wired and wireless controllers; for example, wired connections often reach 117 MB/s while Wi-Fi may struggle at lower speeds due to packet loss.
Minimal Overhead: It typically maintains low CPU utilization (often below 5–10%) during tests.
💡 Pro Tip: If you find LANBench too basic, professional reviewers often supplement it with tools like iPerf for more tunable options or TamoSoft Throughput Test for graphical output. If you'd like, I can: Help you find the download link for LANBench Compare it to more modern tools like iPerf3 Explain how to interpret your results if they seem slow and validate configuration or hardware changes.
LANBench will output critical metrics that hardware-only benchmarks ignore:
| Metric | What it measures | Good Threshold (LAN) | | :--- | :--- | :--- | | TTFT (Time To First Token) | Latency from request send to first token back. | < 100ms for streaming. | | Token/s (throughput) | Tokens generated per second across the network. | > 80% of local speed. | | P95 Latency | Worst-case latency for 95% of requests. | < 500ms for interactive use. | | Request Failures | Timeouts or connection resets. | 0% on a healthy LAN. |
Before LANBench gained traction, engineers relied on tools like text-generation-webui’s built-in stats or llama.cpp’s --benchmark flag. These tools have a fatal flaw: they measure internal compute speed.
Consider this scenario:
Standard benchmarks would report "100 t/s." LANBench would report the truth: "65 t/s due to network jitter and JSON parsing."
LANBench is a benchmarking tool designed to measure and analyze the performance of Local Area Networks (LANs). It evaluates throughput, latency, packet loss, jitter, and protocol behavior under varying traffic patterns to help network engineers quantify performance, identify bottlenecks, and validate configuration or hardware changes.