- Home
- PC benchmark tools
- LMbench
LMbench
What it is
LMbench is a free, self-hosted benchmark suite for measuring data movement and system performance. Its workloads cover latency and bandwidth, processor operations, memory hierarchy, cache and TLB behavior, disks, file systems, networks, pipes, virtual memory, and process activity. Results can report time and CPU cycles per operation for latency, and megabytes per second and CPU cycles per megabyte for bandwidth. Other listed measurements include null process execution time, context switching, network transfers, socket operations, and large-transfer throughput. The command-line README provides benchmark commands and a `make see` option for comparing results with system data included in the repository. The documentation names UNIX/POSIX, Linux, FreeBSD, NetBSD, Solaris, SunOS, AIX, HP-UX, IRIX, OSF/1, and partial Windows NT support; Linux and Windows are listed as platforms. The project is GPL-2.0 licensed. Its documentation cautions that other system activity, particularly CPU cache and cycle effects, can influence results and recommends repeated runs with output comparisons.
For and against
For
- Covers latency, bandwidth, memory, and networking.
- Includes comparison with repository system results.
- Offers command-line benchmarking and automation.
- GPL-2.0 licensed and free to use.
Against
- Windows NT support is partial.
- Other system activity can affect measurements.
- Documentation recommends multiple runs for comparison.
The verdict
LMbench provides broad system benchmarks and a way to compare results with data in its repository. Measurements can be affected by system activity, so its documentation recommends repeating runs and comparing outputs.
Who it suits
LMbench may suit users who need to measure system performance across memory, storage, process, and network workloads. It is a self-hosted command-line tool with named support for Unix-like systems and partial Windows NT support.
Plans and prices
lmbench
GPL-2.0 licensed benchmark software
Read 9 Oct 2026 · github.com
On the record
- Areas measured
- Listed benchmark areas include memory, disks, file systems, networks, pipes, and virtual memory systems.github.com · read 8 Oct 2026
- Benchmark coverage
- The change log lists benchmarks for cache-line size, effective TLB size, cache hierarchy, operation latency, memory bandwidth, and SysV semaphore latency.github.com · read 9 Oct 2026
- Benchmark focus
- The benchmark documentation describes its theme as measuring data movement and its cost.github.com · read 8 Oct 2026
- CLI or automation
- Yesgithub.com · read 23 Sep 2026
- Compare results
- The README says `make see` displays results compared with the other system results included in the repository.github.com · read 9 Oct 2026
- Comparison database
- Yesgithub.com · read 23 Sep 2026
- Included comparison data
- The README refers to other system results included in the repository and documents listing results by operating system.github.com · read 9 Oct 2026
- Latency and bandwidth
- It lists time and CPU cycles per operation for latency, and MB per second and CPU cycles per MB for bandwidth.github.com · read 8 Oct 2026
- License
- The repository identifies its license as GPL-2.0.github.com · read 9 Oct 2026
- Linux distributions
- The README mentions Euler and CentOS and says to uncomment lines 250 and 251 for those operating systems.github.com · read 9 Oct 2026
- LoongArch
- The README gives `make results OS=mips64el` as the command to run the benchmark on LoongArch.github.com · read 9 Oct 2026
- Maker information
- The repository is under Intel's GitHub organization; the repository page provides no maker description or website.github.com · read 9 Oct 2026
- Measurement caveat
- The README warns that benchmarks can be affected by other activity on the system, particularly CPU cache and CPU cycle effects.github.com · read 8 Oct 2026
- Memory testing
- The change log says the `stream` benchmark reports memory bandwidth using kernels from John McCalpin's STREAM and STREAM version 2 benchmarks.github.com · read 9 Oct 2026
- Network measurements
- Network measurements listed include transfers per second, CPU cycles per transfer, socket/bind/close per second, and throughput for large transfers.github.com · read 8 Oct 2026
- Operating systems
- UNIX/POSIX; Linux; FreeBSD; NetBSD; Solaris; SunOS; AIX; HP-UX; IRIX; OSF/1; Windows NT (partial)github.com · read 23 Sep 2026
- Platform references
- The benchmark documentation lists all Unix systems and Windows NT among its systems, and the README gives a LoongArch run command using OS=mips64el.github.com · read 8 Oct 2026
- Process measurements
- The benchmark list includes null process execution time and context switching.github.com · read 8 Oct 2026