On Thursday, 5 December 2013 at 08:21:07 UTC, Jacob Carlborg
wrote:
On 2013-12-04 17:20, Dicebot wrote:
Ah this is rather sad. It makes tests results somewhat
unstable because
client load interferes with server load. Ideally client should
be
separate machine and has more powerful h/w than server.
However, that
also requires ~gigabit network in between as well as matching
network
cards to hit the limits of top server performance - this makes
such
tests rather cumbersome to execute.
At the very least you should use process affinity to separate
resources
between client and server in more predictable manner.
Isn't it most important that all languages were tested in the
same way?
Depends on your goals :) If you want just to say "hah, your
language is nothing compared to my language" it is enough. If you
want to make some observations/conclusions about specific
implementation performance and how it scale for different
conditions it becomes important to remove as many side impact as
possible. And of course at high load/concurrency levels client
competing with server for connections does make notable impact.
In other words, it is good enough for comparison but not good
enough for actual performance analysis.
