Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 6 Nov 2012 09:46:22 -0800
From:      Garrett Cooper <yanegomi@gmail.com>
To:        Wojciech Puchar <wojtek@wojtek.tensor.gdynia.pl>
Cc:        Olivier Smedts <olivier@gid0.org>, Yuri <yuri@rawbw.com>, "freebsd-hackers@freebsd.org" <freebsd-hackers@freebsd.org>
Subject:   Re: pgbench performance is lagging compared to Linux and DragonflyBSD?
Message-ID:  <B4315143-CFCF-4B58-8E8A-394F1BCC21C2@gmail.com>
In-Reply-To: <alpine.BSF.2.00.1211061752340.20322@wojtek.tensor.gdynia.pl>
References:  <50980ADD.4010402@rawbw.com> <alpine.BSF.2.00.1211061016110.18204@wojtek.tensor.gdynia.pl> <CABzXLYPYaVuEFc2SEY1H2Wa0T6A_SHuTu=W3UEZ554j5BR01bQ@mail.gmail.com> <alpine.BSF.2.00.1211061752340.20322@wojtek.tensor.gdynia.pl>

next in thread | previous in thread | raw e-mail | index | archive | help
On Nov 6, 2012, at 8:55 AM, Wojciech Puchar <wojtek@wojtek.tensor.gdynia.pl>=
 wrote:

>>> Tuning operating system for single benchmark is an example of that child=
ish
>>> behaviour.
>>=20
>> LOL. That's what "we" did several years ago :
>> http://people.freebsd.org/~kris/scaling/dfly.html
>=20
> i've seen that page some time ago but i don't really care of it.
> i just wasn't interested.
>=20
> Still - DOING such benchmark is good, as it can show general problems in u=
sed algorithms.
>=20
> But working on software to make it better in some kind of synthetic benchm=
ark is common in commercial software world. ("We have more performance per b=
uck than company X")

"Synthetic benchmarks" as you put it shouldn't be the ultimate basis for a d=
ecision, but instead allow users to gauge whether or not a certain software o=
r hardware configuration is suitable for their given workload. No more, no l=
ess. The fact that they're being used in this manner is a bit like a salesma=
n selling snake oil as the results aren't necessarily the result of a "best"=
 configuration for all competing platforms, but instead an unknown configura=
tion in this case.

A similar statement about the importance of micro benchmarks can be made...

Thanks,
-Garrett=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?B4315143-CFCF-4B58-8E8A-394F1BCC21C2>