4 concurrent kernel builds w/ make -j (on three disks)
4 concurrent nfs reads as client from Suns under Solaris 2.6
4 concurrent nfs reads as server w/ unfs-daemons running under
realtime scheduling policy
mpg123 playing my favourite CD,
Some sucker doing a recursive "Visual Schnauzer"-update under xv,
2100 incoming emails running into an innd mail2news gw,
88 open network connections, 68 of them are fwd-X11 from sshd (lo),
one tcp-connection which eats up the rest of the network bandwith with
dd if=/dev/zero|sock remotemachine:10001
mbone tunneling/routing w/ mrouted (sucks, cause pruning is defective)
heavy syslog activity from nfs clients comlaining and tcplog logging
the smtp and nntp connections
4 netscapes from 6 interactive users
1 daily cron with updatedb
1 adsm backup with SCO binary
All together up to 431 processes, max 121 running, while I write this
mail. At the load of 70, it started to get slightly sluggish (but yet
perfectly responsive), and the music wasn't clean any more. :-( But,
hey, hardware is P150, 128 MB EDO 60 ns, 2xNCR810A, 4 disks, 1 IDE,
tulip, SB32, Matrox Mystique 4MB. 2x100 MB swap at same prio,
libc-5.4.46, gcc-2.7.2.1.
The test may not be very representative, but it shows that this kernel
is dummy-proof.
As I told them that they can't buy it, they turned grey. Then I showed
them my U-1-167 under Sol. 2.6 at 90% CPU load displaying the mess
from the linux box... The Sun started paging because of the nfs server
load, caused by the test.
Roman.
Computer Center University of Freiburg, Germany.
"The whole world is about three drinks behind." (Humphrey Bogart)
-
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/