Here's an attempt to create yet another damn Linux server benchmarking script.
This script isn't an attempt to be a golden standard. It's just yet another bench script to add to your arsenal. Included are several tests that I think are most beneficial for the end-user. If there's features that you would like to see added, feel free to submit an issue describing your feature request or fork the project!
View YABS usage stats here.
curl -sL yabs.sh | bash
This script has been tested on the following Linux distributions: CentOS 6+, Debian 8+, Fedora 30, and Ubuntu 16.04+. It is designed to not require any external dependencies to be installed nor elevated privileges to run.
Local fio/iperf3 Packages: If the tested system has fio and/or iperf3 already installed, the local package will take precedence over the precompiled binary.
IPv6-Only Machines: The above command will not work on IPv6-only machines. See below.
High Bandwidth Usage Notice: By default, this script will perform many iperf network tests, which will try to max out the network port for ~20s per location (10s in each direction). Low-bandwidth servers (such as a NAT VPS) should consider running this script with the -r flag (for reduced iperf locations) or the -i flag (to disable network tests entirely).
By default, the script runs all three tests described in the next section below. In the event that you wish to skip one or more of the tests, use the commands below:
curl -sL yabs.sh | bash -s -- -{fdighr49}
-f/-dthis option disables the fio (disk performance) test-ithis option disables the iperf (network performance) test-gthis option disables the Geekbench (system performance) test-hthis option prints the help message with usage, flags detected, and local package (fio/iperf) status-rthis option reduces the number of iperf locations (Online.net/WorldStream/HE.net) to lessen bandwidth usage-4this option overrides the Geekbench 5 performance test and runs a Geekbench 4 test instead-9this option runs the Geekbench4 test in addition to the Geekbench 5 test
Options can be grouped together to skip multiple tests, i.e. -fg to skip the disk and system performance tests (effectively only testing network performance).
- fio - the most comprehensive I/O testing software available, fio grants the ability to evaluate disk performance in a variety of methods with a variety of options. Four random read and write fio disk tests are conducted as part of this script with 4k, 64k, 512k, and 1m block sizes. The tests are designed to evaluate disk throughput in near-real world (using random) scenarios with a 50/50 split (50% reads and 50% writes per test).
- iperf3 - the industry standard for testing download and upload speeds to various locations. This script utilizes iperf3 with 8 parallel threads and tests both download and upload speeds. If an iperf server is busy after 10 tries, the speed test for that location/direction is skipped.
- Geekbench - Geekbench is a benchmarking program that measures system performance, which is widely used in the tech community. The web URL is displayed to be able to see complete test and individual benchmark results and allow comparison to other geekbench'd systems. The claim URL to add the Geekbench result to your Geekbench profile is written to a file in the directory that this script is executed from. By default, Geekbench 5 is the only Geekbench test performed, however, Geekbench 4 can also be toggled on by passing the appropriate flag.
This script relies on external binaries in order to complete the performance tests. The network (iperf3) and disk (fio) tests use binaries that are compiled by myself utilizing a Holy Build Box compiliation environment to ensure binary portability. The reasons for doing this include ensuring standardized (parsable) output, allowing support of both 32-bit and 64-bit architectures, bypassing the need for prerequisites to be compiled and/or installed, among other reasons. For the system test, a Geekbench tarball is downloaded, extracted, and the resulting binary is run. Use this script at your own risk as you would with any script publicly available on the net. Additional information regarding the binaries, including compilation notes and steps, can be found in the bin directory's README page.
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
# Yet-Another-Bench-Script #
# v2020-02-04 #
# https://github.com/masonr/yet-another-bench-script #
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
Tue Feb 4 19:04:24 UTC 2020
Basic System Information:
---------------------------------
Processor : Intel(R) Xeon(R) CPU E3-1270 v6 @ 3.80GHz
CPU cores : 8 @ 800.098 MHz
AES-NI : ✔ Enabled
VM-x/AMD-V : ✔ Enabled
RAM : 31G
Swap : 0B
Disk : 221G
fio Disk Speed Tests (Mixed R/W 50/50):
---------------------------------
Block Size | 4kb (IOPS) | 64kb (IOPS)
------ | --- ---- | ---- ----
Read | 69.37 MB/s (17.3k) | 106.51 MB/s (1.6k)
Write | 69.57 MB/s (17.3k) | 107.07 MB/s (1.6k)
Total | 138.94 MB/s (34.7k) | 213.59 MB/s (3.3k)
| |
Block Size | 512kb (IOPS) | 1mb (IOPS)
------ | ----- ---- | --- ----
Read | 133.52 MB/s (260) | 141.90 MB/s (138)
Write | 140.61 MB/s (274) | 151.35 MB/s (147)
Total | 274.13 MB/s (534) | 293.26 MB/s (285)
iperf3 Network Speed Tests (IPv4):
---------------------------------
Provider | Location (Link) | Send Speed | Recv Speed
| | |
Bouygues Telecom | Paris, FR (10G) | 1.41 Gbits/sec | 1.13 Gbits/sec
Online.net | Paris, FR (10G) | 1.44 Gbits/sec | 1.29 Gbits/sec
Worldstream | The Netherlands (10G) | 1.18 Gbits/sec | 1.22 Gbits/sec
wilhelm.tel | Hamburg, DE (10G) | 805 Mbits/sec | 1.13 Gbits/sec
Biznet | Bogor, Indonesia (1G) | 768 Mbits/sec | 38.5 Mbits/sec
Hostkey | Moscow, RU (1G) | 503 Mbits/sec | 686 Mbits/sec
Velocity Online | Tallahassee, FL, US (10G) | 2.74 Gbits/sec | 2.67 Gbits/sec
Airstream Communications | Eau Claire, WI, US (10G) | 3.36 Gbits/sec | 963 Mbits/sec
Hurricane Electric | Fremont, CA, US (10G) | 6.34 Gbits/sec | 3.76 Gbits/sec
iperf3 Network Speed Tests (IPv6):
---------------------------------
Provider | Location (Link) | Send Speed | Recv Speed
| | |
Bouygues Telecom | Paris, FR (10G) | 1.44 Gbits/sec | 1.25 Gbits/sec
Online.net | Paris, FR (10G) | 1.36 Gbits/sec | 972 Mbits/sec
Worldstream | The Netherlands (10G) | 1.19 Gbits/sec | 1.20 Gbits/sec
wilhelm.tel | Hamburg, DE (10G) | 826 Mbits/sec | 1.14 Gbits/sec
Airstream Communications | Eau Claire, WI, US (10G) | busy | busy
Hurricane Electric | Fremont, CA, US (10G) | 6.36 Gbits/sec | 2.95 Gbits/sec
Geekbench 4 Benchmark Test:
---------------------------------
Test | Value
|
Single Core | 5587
Multi Core | 19093
Full Test | https://browser.geekbench.com/v4/cpu/15200550
GitHub's CDN does not resolve via IPv6. You will need to run the following command to download and run the script.
curl -s -k -g --header 'Host: raw.githubusercontent.com' https://[2a04:4e42::133]/masonr/yet-another-bench-script/master/yabs.sh | bash
(2a04:4e42::133 is fastly.net's [GitHub's CDN Provider] IPv6 address)
This script was inspired by several great benchmarking scripts out there, including, but not limited to, bench.sh, nench.sh, ServerBench, among others. Members of the HostedTalk, LowEndSpirit, and LowEndTalk hosting-related communities play a pivotal role in testing, evaluating, and shaping this script as it matures.
DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE
Version 2, December 2004
Copyright (C) 2019 Mason Rowe <[email protected]>
Everyone is permitted to copy and distribute verbatim or modified
copies of this license document, and changing it is allowed as long
as the name is changed.
DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE
TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
0. You just DO WHAT THE FUCK YOU WANT TO.