Slashdot Mirror
Benchmarking the Scalability of BSD and Linux
Fefe writes "I recently did some benchmarks for a talk about scalable network programming I held at Linux Kongress 2003. The benchmark results turned out to be surprising enough to present them on their own. This ought to end those pesky flame wars about whose IP stack or memory management scales better. Or maybe not."
He was running unstable versions of the other systems too (except NetBSD), though. Imagine if he had run the comparison with Linux 2.6 against FreeBSD 4. If it was trounced in the tests, which given the way 5 performed and the major improvements in that version it ought to have been, all of the BSD zealots would be whinging about how comparing the development linux kernel to the antiquarian stable FreeBSD kernel is not a fair test. This way, there is nothing really to whine about.
From the article:
The OpenBSD and FreeBSD graphs stop early because OpenBSD crashed when I forked more processes, and I couldn't find out how to increase FreeBSD's system limit on the number of processes (sysctl said the value was read-only).
MOD THIS STORY DOWN AS A TROLL! BSD NOT DEAD!
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
>Nope, kern.maxproc is really read-only in a /boot/loader.conf (which doesn't
>running system even in securelevel -1. You have
>to set it in
>seem to be prominently documented anywhere, so >not finding it is nothing to blame fefe for).
It is something to blame him for. He should have asked the advice of someone familiar with the project. Just because it is not in the documentation is no reason to ask support. Especially in a case like this which is a high profile comparison between well known O/S's.