Skip to main content

The only benchmark tool on the block crazy enough to abuse the UNIX yes command to test your CPU

Project description

yesBenchMark

PyPI version License PyPI Downloads GitHub downloads

yesBenchMark is a crazy tool for benchmarking the CPU in UNIX & UNIX-like systems. It abuses the yes command to load the CPU and measures CPU utilization, speed, and compute.

yesBenchMark demo (animated)

Features

  • Extremely tiny & light
  • Can run on extremely old hardware (OS X 10.4 Tiger is oldest it can run on of all the stuff out there, with a modest G3)
  • Can do a ton of different modes, varying in load from light to extremely punishing to see how far you can push your CPU:
    • time - See how much the CPU can output in 30 & 60 seconds respectively in a single thread (yes)
    • multicore - Multi-core version of the time benchmark
    • freq - Bench how fast your CPU core can turbo
    • multi-freq - Multi-core version of freq benchmark
    • stress - See how little of the total CPU a single thread can use, also includes a stability measurement
    • compute - Check how long it takes the CPU to compute 27,000,000,000 hunks of data (each hunk is a single y\n)
    • compute-xt - compute benchmark but with 10x the hunks
  • Written in Python & extremely portable, so runs almost anywhere
  • Probably one of the most practical benchmarks out there - measures CPU load, CPU frequency, and other things encouuntered in day-to-day workloads; none of the shenanigans that most tools use to grade your CPU's performance

Usage

Usage: yesbenchmark [options]
Options:
  --help, -h          Show this help message
  --mode=MODE         Run specific mode of benchmark

Modes:
all                   Run all modes of benchmarks
time                  Run time-bound benchmark
multicore             Run time-bound benchmark for all cores
freq                  Run frequency-bound benchmark
multi-freq            Run multi-core version of frequency-bound benchmark
stress                Run stress/utilization benchmark
compute               Run compute benchmark (time taken depends on CPU speed)
compute-xt            Run a more extreme version of compute benchmark

Well that is...all the capabilities of the program. We recommend the compute mode for most people.

Scores for reference (performance is compute & extreme is compute-xt):

M1 Max with 8 P-cores & 2 E-cores (10 cores total):

  • P70,363 → performance mode
  • X685 → extreme mode

i7-2600 with 8 cores:

  • P55,609 → performance mode
  • X557 → extreme mode

G5 with 1 core:

  • P572 → performance mode
  • {untested} → extreme mode (would take forever on my iMac G5 and overheat it)

Note: Do NOT run the program with zero arguments unless you want to run ALL benchmarks except the compute-xt one. This will take a long time and likely overheat your computer.

Installation

Requirements:

  • Python 3.8 or later (we don't test for older, probably works down to 3.6)
  • OS X 10.5 Leopard or later (if too old to be supported by Homebrew, please use Tigerbrew to install Python 3.10) or Linux 3.2 or later (BSDs may be supported at a later time when I have time to spin up one in a virtual machine)
    • Other UNIXes are probably going to work, so long as they are at least UNIX-like. We don't currently test for these other systems, but fixes are welcome. We might also start testing for BSD sometime.

To install yesBenchMark via pip:

pip install yesbenchmark

Or if you have cloned the repo for development (note that this will also allow you to edit the src code without re-installing for ease):

pip install -e .

Compiling and installing as a binary

Requirements:

  • The above requirements
  • If macOS, an Intel x86_64 CPU or better
  • MacOSX 10.9 or later if on Mac

To compile:

python3 compile.py

After that, just copy the binary in dist out and into a directory in your PATH.

License

This project is licensed under the GNU General Public License v3.0 (GPLv3). Please note that it will also be compatible with any future versions of this license, should they be released. All past and future versions of this software are covered by this license. See the LICENSE file for full details.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

yesbenchmark-1.0.0.post2.tar.gz (23.6 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

yesbenchmark-1.0.0.post2-py3-none-any.whl (24.5 kB view details)

Uploaded Python 3

File details

Details for the file yesbenchmark-1.0.0.post2.tar.gz.

File metadata

  • Download URL: yesbenchmark-1.0.0.post2.tar.gz
  • Upload date:
  • Size: 23.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.11

File hashes

Hashes for yesbenchmark-1.0.0.post2.tar.gz
Algorithm Hash digest
SHA256 b0cfd543021fb923616996ccb7dd71dfdca3c013b37e60740507a44691933052
MD5 9e8b51ba3c591a115e472a0d18903bf8
BLAKE2b-256 322d8977027eff4ef36bb22b4a36b0250acc7bc9089f465860052691ab60132d

See more details on using hashes here.

File details

Details for the file yesbenchmark-1.0.0.post2-py3-none-any.whl.

File metadata

File hashes

Hashes for yesbenchmark-1.0.0.post2-py3-none-any.whl
Algorithm Hash digest
SHA256 4c4bc2abacc994b89a2b474de174ac326d3eddce233b02bd8d921cae8bded55f
MD5 ffa3c079d63de870aabdcf78d7822048
BLAKE2b-256 75caf440b989d8acaaf2c7fcda68e23fd81d80d3e5185c3088f5fc818ed6eb9b

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page