loosely based on the established freevps.us/bench.sh
includes CPU and ioping measurements
reduced number of speedtests (9 x 100 MB), while retaining useful European and North American POPs
runs IPv6 speedtest by default (if the server has IPv6 connectivity)
has a 50-second timeout for each speedtest, so you don't end up waiting 10 minutes for that one slow speedtest from halfway around the globe (this means that any speedtest result < 2 MB/s will be squelched)

The script was originally intended to be used only by me, so I didn't put much effort into ensuring safety, security, and interoperability.
I welcome any improvements, just send me a pull request.

Disclaimer

You've probably noticed that the usage examples below have you directly run a script from an unauthenticated source (as so many "easy-install" and benchmark scripts do).

I didn't think I'd have to mention that this is a potential security risk — really, if you're at the point where you're benchmarking Linux VMs, I would assume you know how much harm a rogue shell script could potentially do to your system…

What's more, nench.sh downloads a statically built binary to run the IO latency tests. I assure you it is and always will be a clean unmodified build of ioping, but how do you know you can trust me?

So, basically: use nench.sh at your own risk, and preferably not on production systems (which is a bad idea anyway, because it will hammer your harddisk and network for up to several minutes).

Usage example

(curl -s wget.racing/nench.sh | bash; curl -s wget.racing/nench.sh | bash) 2>&1 | tee nench.log

(wget -qO- wget.racing/nench.sh | bash; wget -qO- wget.racing/nench.sh | bash) 2>&1 | tee nench.log