by v.nagasriniv » Thu, 04 Oct 2001 20:19:23
How is it (NetBSD) better than linux 2.4.x kernels
when comparing Networking features ( routing, etc..) and Porting to
new hardware, Considering inview, thinking for a router with say IXP1200
processor..
Can any give, a complete view...
thanks,
srinivas.
by bbal » Fri, 19 Oct 2001 08:02:16
i'll give one example:
- i use NetBSD on an IBM WorkPad Z50 (MIPS CPU, hpcmips port of NetBSD)
- NetBSD recognizes and uses all installed RAM
- Linux/CE does not
- to be fair, on the other hand, NetBSD does not support audio on the Z50,
while Linux/CE does
my pov...
Quote:>HI,
> How is it (NetBSD) better than linux 2.4.x kernels
>when comparing Networking features ( routing, etc..) and Porting to
>new hardware, Considering inview, thinking for a router with say IXP1200
>processor..
>Can any give, a complete view...
>thanks,
>srinivas.
1. Linux problem with compiles over NFS, help needed, NetBSD better?
I have been trying to run software compiles here in our environment
which is heavily NFS cross-mounted. Our main NFS server is a Sparc 1+
running SunOS 4.1.3.
My local machine is a Pentium 120 running Linux 1.2.13.
My problem is that compiles are painfully slow when running on the
Linux platform while reading/writing from the NFS mounted volumes. If
everything is local, then there is no speed problem at all (it's
awesome!).
Here is some data that I've collected. Times are all for the same
compile:
local disk, local gcc:
8.88user 5.43system 0:22.93elapsed 62%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
local disk, remote gcc:
8.62user 7.18system 1:06.06elapsed 23%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
remote disk, local gcc:
8.87user 9.80system 7:43.72elapsed 4%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
remote disk, remote gcc:
8.93user 11.30system 10:44.28elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
Moving the source code or binary destination remote or local
(separately, to determine if it is reading or writing that's the
problem):
local disk (src), remote disk (Lib,Bin), local gcc:
8.58user 6.61system 0:43.96elapsed 34%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
remote disk (src), local disk (Lib,Bin), local gcc:
8.89user 9.42system 7:20.34elapsed 4%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+0minor)pagefaults 0swaps
Does anyone have any idea what is causing this slowness? It is so
slow as to be unusable. Sparcs compile much faster.
Is there some NFS tweaking that I need to do?
Is NetBSD a better platform for heavy NFS usage?
Suggestions welcome, solutions even more welcome...
;-)
Thanks.
--glen
--
**************************************************************
Midnight Networks Inc. * 200 Fifth Avenue * Waltham, MA 02154
Glen B. Glater Principal, Strategic Networking Group
Phone: (617) 890-1001 Fax: (617) 890-0028
Midnight Networks Inc. WWW page http://www.midnight.com
2. "no COMMENT variable dfined for this port"
3. Am I better off switching to another video card--is the agp bus the problem
5. What is better solaris or netbsd
6. Are there office apps for linux?
7. Am I missing some news or did NetBSD 1.1 not get announced?
9. freebsd, netbsd which is better?
10. Is AMD better or Intel better for Linux?
11. Linux NOT Enduser -> Better Linux -> Next Linux ?
12. This clone thing...am I stupid, or am I right?