Messages in this thread | | | Date | Fri, 24 Dec 1999 10:43:38 -0000 (GMT) | From | Anton Ivanov <> | Subject | Re: Bloat? (khttpd) |
| |
-----BEGIN PGP SIGNED MESSAGE-----
On 23-Dec-1999 Dan Hollis wrote: > On Thu, 23 Dec 1999, Stephen Frost wrote: >> I believe the duty to produce numbers goes to the person trying to get >> the code included. The idea being that you have to prove the solution is >> useful >> w/ code and numbers before it will be included in the kernel. At least, >> that's >> the way I always saw it. > > http://www.fenrus.demon.nl/performance.html
Interesting results. Did Apache use mmap? Or it was plain old file IO?
This is getting OT, but I hope that I will be excused askining a few Qs:
1. What is the difference between these tests and the c't replication of the Mindcraft bench. C't showed that properly configured apache in a non-MS oriented network setup has more than reasonable performance. These results contradict. Where is the difference that causes it?
2. All this discussion seems to go back to the Mindcraft scenario. As noted in (1) C't showed linux to perform decently with one network card and lag when put in windows favouring multiether setup. From the described setup by C't (and Mindcraft): either their descriptions were incomplete, or normal unix put in those conditions will not load balance neatly. So nothing amazing there...
3. Assuming what was said on the list about CPU usage for: - - parsing headers (high) - - transfering files (low) - - benchmarks (questions, questions ...)
Does it really worth it to go the Mindcraft/MS way and hastily put a httpd implementation in the kernel?
Maybe a detailed analysis of C't results (as being less biased) and some numbers first???
Thanks, And once again pls execuse me for the OT.
- ---------------------------------- Anton R. Ivanov IP Engineer Level3 Communications RIPE: ARI2-RIPE E-Mail: Anton Ivanov <aivanov@eu.level3.net> - ---------------------------------- -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.0 (GNU/Linux) Comment: For info see http://www.gnupg.org
iQEVAwUBOGNOWSlWAw/bM84zAQHLEQf9GM4WAtHCunkL6hMO3XYE0+cmSM4Qob42 6TtgvDdXdxpPJv26rgntnpNLZAc9AccVuzjvNebypv8sYJReKHAiHU1LGiZTxgvf 30YpYx1zBOpbvOHcjYfqaSISk+jVAokm4WQ48EWtZWXgXXqaSkyI1UDTL2bHmNDE okqHy+say2XG3qjuyrlmAcr+l4uFYQdhGYmUbT2sUHo1cRwU9bsl3UkoZjMu/rpM K/oCkqJ8eGu0ZC7eiVr1cD57TCFzP5wsK0ji3qTaalIN5Wcx7MbkQQiZLA5a5oVw 7exB3oUvgHNkzPzEU0ylb1w+AjjzBAQb1xTZfXkMWIMMCftjSayIBQ== =groK -----END PGP SIGNATURE-----
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |