On Fri, 2004-07-16 at 11:15, Mark Knecht wrote:
Florin Andrei wrote:
The reality is quite the opposite. With IO-APIC the IRQs suffer from
more overlapping than without it. Which is kind of strange to me.
BTW, the mobo is based on the NForce v1 chipset.
As per Clemens' note this is complicated. I'm not sure I understand your
feedback about 'suffers from more overlapping'.
More devices were crowded upon the same IRQ.
Could you post the
output of /proc/interrupts for the same machine configuration with
IO-APIC and old style interrupts?
Well, i had to make some changes anyway, so my NForce mobo has been
officially demoted to the status of Internet firewall mobo. Now my PC's
mobo is SiS based, same generation as NForce2 and probably pretty much
the same performance.
There's a distasteful overlap between EMU10K1 and eth0 but i'm not using
the network too much while doing audio work:
[florin@rivendell florin]$ cat /proc/interrupts
CPU0
0: 28956075 IO-APIC-edge timer
1: 10 IO-APIC-edge i8042
8: 1 IO-APIC-edge rtc
9: 0 IO-APIC-level acpi
12: 58 IO-APIC-edge i8042
14: 35556 IO-APIC-edge ide0
15: 33 IO-APIC-edge ide1
16: 2503314 IO-APIC-level ohci1394, nvidia
17: 88 IO-APIC-level ide2, ide3
19: 3098 IO-APIC-level eth0, EMU10K1
20: 0 IO-APIC-level ohci_hcd
21: 0 IO-APIC-level ohci_hcd
23: 0 IO-APIC-level ehci_hcd
NMI: 0
LOC: 28957230
ERR: 0
MIS: 0
The kernel is the latest Fedora 2 update (2.6.6-1.435.2.3) rebuilt by
myself with pre-empt and IO-APIC enabled.
--
Florin Andrei
http://florin.myip.org/