I noticed that A. all of the interrupts for the ethernet card are being sent to the first virtual processor only, and are not being shared among the CPUs, and B. the number of interrupts generated by eth0 is rather high. The system has an Intel e1000 network card and is running Debian Testing, using kernel 2.6.8-2-686-smp (there is a 2.6.10 update I'm holding back until I can test it).
CODE
# cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3
0: 1645265455 3423810505 1645306700 1645291045 IO-APIC-edge timer
9: 0 0 0 0 IO-APIC-level acpi
15: 27 1 0 0 IO-APIC-edge ide1
177: 1048805038 0 0 0 IO-APIC-level eth0
185: 9561 40447961 53343831 157 IO-APIC-level 3ware Storage Controller
NMI: 0 0 0 0
LOC: 4065107265 4065107264 4065107260 4065107262
ERR: 0
MIS: 0
CPU0 CPU1 CPU2 CPU3
0: 1645265455 3423810505 1645306700 1645291045 IO-APIC-edge timer
9: 0 0 0 0 IO-APIC-level acpi
15: 27 1 0 0 IO-APIC-edge ide1
177: 1048805038 0 0 0 IO-APIC-level eth0
185: 9561 40447961 53343831 157 IO-APIC-level 3ware Storage Controller
NMI: 0 0 0 0
LOC: 4065107265 4065107264 4065107260 4065107262
ERR: 0
MIS: 0
CODE
# cat /proc/net/dev
Inter-| Receive | Transmit
face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed
lo:661406885 8751485 0 0 0 0 0 0 661406885 8751485 0 0 0 0 0 0
eth0:3535330893 626929625 0 0 0 0 0 0 743467137 860537996 0 0 0 0 0 0
sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Inter-| Receive | Transmit
face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed
lo:661406885 8751485 0 0 0 0 0 0 661406885 8751485 0 0 0 0 0 0
eth0:3535330893 626929625 0 0 0 0 0 0 743467137 860537996 0 0 0 0 0 0
sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Any thoughts?