Re: [Qemu-discuss] Standard benchmarks for VMs

[Phil Ehrens]

I use this toy benchmarking script:

#!/bin/sh
start=$(date +%s%N)
dd if=/dev/zero of=100M.zero count=200000 2>/dev/null
gzip 100M.zero
sync
stop=$(date +%s%N)
rm -f 100M.zero*
total=$(echo "$stop - $start" | bc)
echo "$total nanoseconds"

That 'sync' command really makes the VM look bad,
compared to the same script without the 'sync' ;^)

Tony Su wrote:
> I guess it really depends on what you're trying measure/benchmark.
> 
> Benchmarking is like accounting...
> I've seen benchmarks that "prove" all sorts of things, not always
> relevant to the real world and certainly questionable relevance to
> various setups.
> 
> So, the benchmark you reference looks like an interesting and
> reasonable benchmark, storage performance is an important factor in
> comparing and optimizing overall performance. But, how you use it
> might still be important.
> 
> IMO,
> Tony
> 
> On Tue, Jul 30, 2013 at 3:03 AM, Gabriele Paciucci <address@hidden> wrote:
> > I'm new for this ml, sorry if this mail is not in the right place!!!
> >
> > I'm looking for standard benchmarks to test VMs.
> > I have seen several presentations from IBM and Red Hat using FFSB. Is there 
> > a "widely used" configuration file for FFSB? Just to have a starting point?
> >
> > Thanks in advance.
>