You are not logged in.
Currently iam using a custom build kernel 2.6.21 with viper patchset on an Sony Vaio C2Z Core2Duo T5500.
Actually i disabled Kernel IRQ Balancing in Kernel config before build and use the irqbalance deamon instead.
Anyway it seems like using Kernel balance or the deamon doesnt make any difference. I also googled around
a bit and try to get more information about balancing technics in general, but didnt find any real useful informations at all.
Now i just wanted to ask what are your experiences and suggestions.
Offline
I have tried both the daemon and enabling irqbalance in the kernel config (arch default kernel config disabled). I found that irqbalancing with the kernel module did not work on my machine. cat /proc/interrupts still showed all the interrupts going to cpu0. With the daemon the interrupts were divided between the two cpu's. However I found that performance seemed to be degraded with irqbalance active, sound etc used to skip during heavy use of the machine. With irqbalance inactive no sound skipping etc. I would suggest not using irqbalance at all.
---for there is nothing either good or bad, but only thinking makes it so....
Hamlet, W Shakespeare
Offline
My proc table is as follow
with irqbalance daemon
[root@seli-vaio ikarus]# cat /proc/interrupts
CPU0 CPU1
0: 30027 0 IO-APIC-edge timer
1: 132 0 IO-APIC-edge i8042
9: 1282 0 IO-APIC-fasteoi acpi
12: 120 0 IO-APIC-edge i8042
14: 1377 0 IO-APIC-edge libata
15: 0 0 IO-APIC-edge libata
16: 0 225322 IO-APIC-fasteoi uhci_hcd:usb4, eth0, ohci1394, yenta, nvidia
17: 2507 0 IO-APIC-fasteoi uhci_hcd:usb2, tifm_7xx1
18: 6082 0 IO-APIC-fasteoi uhci_hcd:usb3, ipw3945
20: 6020 0 IO-APIC-fasteoi libata
21: 3 0 IO-APIC-fasteoi uhci_hcd:usb1, ehci_hcd:usb5
22: 2650 0 IO-APIC-fasteoi HDA Intel
NMI: 0 0
LOC: 23363 22114
ERR: 0
MIS: 0without
[root@seli-vaio ikarus]# cat /proc/interrupts
CPU0 CPU1
0: 30027 0 IO-APIC-edge timer
1: 132 0 IO-APIC-edge i8042
9: 1282 0 IO-APIC-fasteoi acpi
12: 120 0 IO-APIC-edge i8042
14: 1377 0 IO-APIC-edge libata
15: 0 0 IO-APIC-edge libata
16: 5787 0 IO-APIC-fasteoi uhci_hcd:usb4, eth0, ohci1394, yenta, nvidia
17: 2507 0 IO-APIC-fasteoi uhci_hcd:usb2, tifm_7xx1
18: 6082 0 IO-APIC-fasteoi uhci_hcd:usb3, ipw3945
20: 6020 0 IO-APIC-fasteoi libata
21: 3 0 IO-APIC-fasteoi uhci_hcd:usb1, ehci_hcd:usb5
22: 2650 0 IO-APIC-fasteoi HDA Intel
NMI: 0 0
LOC: 23363 22114
ERR: 0
MIS: 0Seems with activated balancing it sets network and graphic related things to another cpu what maybe could (did not proof that) affect (?under heavy load?) performance of them. It would be nice if these things were better documentated.
Offline
Without
CPU0 CPU1
0: 299 0 IO-APIC-edge timer
1: 3955 0 IO-APIC-edge i8042
7: 790772 0 IO-APIC-edge parport0
8: 0 0 IO-APIC-edge rtc0
9: 0 0 IO-APIC-fasteoi acpi
12: 406952 0 IO-APIC-edge i8042
14: 184 0 IO-APIC-edge libata
15: 245311 0 IO-APIC-edge libata
16: 207212 0 IO-APIC-fasteoi libata, uhci_hcd:usb4
17: 1140694 0 IO-APIC-fasteoi uhci_hcd:usb2, uhci_hcd:usb5, i915@pci:0000:00:02.0
18: 897087 0 IO-APIC-fasteoi Intel ICH5
19: 8324566 0 IO-APIC-fasteoi uhci_hcd:usb3, eth0
20: 1600 0 IO-APIC-fasteoi ehci_hcd:usb1
NMI: 0 0
LOC: 8855456 8706705
ERR: 0
MIS: 0
with irqbalance
CPU0 CPU1
0: 299 0 IO-APIC-edge timer
1: 3957 130 IO-APIC-edge i8042
7: 790772 0 IO-APIC-edge parport0
8: 0 0 IO-APIC-edge rtc0
9: 0 0 IO-APIC-fasteoi acpi
12: 407933 21011 IO-APIC-edge i8042
14: 184 0 IO-APIC-edge libata
15: 246417 0 IO-APIC-edge libata
16: 207215 14913 IO-APIC-fasteoi libata, uhci_hcd:usb4
17: 1158775 0 IO-APIC-fasteoi uhci_hcd:usb2, uhci_hcd:usb5, i915@pci:0000:00:02.0
18: 912392 0 IO-APIC-fasteoi Intel ICH5
19: 8324586 1270 IO-APIC-fasteoi uhci_hcd:usb3, eth0
20: 1600 0 IO-APIC-fasteoi ehci_hcd:usb1
NMI: 0 0
LOC: 8980952 8845140
ERR: 0
MIS: 0
Immediately had sound skipping when loading konqueror. Don't really notice any other performance degradation or improvement.
---for there is nothing either good or bad, but only thinking makes it so....
Hamlet, W Shakespeare
Offline
I reopen this thread after years, in order to ask if finally irqbalance is needed with newer kernels.
Any opinions?
Mikes on AUR
Offline
My /proc/interrupts:
CPU0 CPU1
0: 72420048 3533433 IO-APIC-edge timer
1: 496374 9452 IO-APIC-edge i8042
8: 989 85 IO-APIC-edge rtc0
9: 230 11 IO-APIC-fasteoi acpi
12: 3761927 24131 IO-APIC-edge i8042
14: 1389908 89177 IO-APIC-edge ata_piix
15: 0 0 IO-APIC-edge ata_piix
16: 3688114 231594 IO-APIC-fasteoi nvidia
17: 26112377 2533029 IO-APIC-fasteoi wlan0, eth0
18: 6 5 IO-APIC-fasteoi mmc0
19: 401 2793 IO-APIC-fasteoi firewire_ohci
20: 2125140 124155 IO-APIC-fasteoi ehci_hcd:usb2, uhci_hcd:usb3, uhci_hcd:usb5
21: 0 0 IO-APIC-fasteoi uhci_hcd:usb4, uhci_hcd:usb6
22: 24 2 IO-APIC-fasteoi ehci_hcd:usb1, uhci_hcd:usb7
28: 1030967 82812 PCI-MSI-edge ahci
29: 2302 56625 PCI-MSI-edge hda_intel
NMI: 0 0 Non-maskable interrupts
LOC: 15908881 41203333 Local timer interrupts
SPU: 0 0 Spurious interrupts
PMI: 0 0 Performance monitoring interrupts
PND: 0 0 Performance pending work
RES: 68365200 48065668 Rescheduling interrupts
CAL: 16070 19758 Function call interrupts
TLB: 212743 217255 TLB shootdowns
TRM: 235974 235941 Thermal event interrupts
THR: 0 0 Threshold APIC interrupts
MCE: 0 0 Machine check exceptions
MCP: 694 696 Machine check polls
ERR: 25
MIS: 0Seems to be pretty well balanced to me.
Not using any daemon.
Offline
I reopen this thread after years, in order to ask if finally irqbalance is needed with newer kernels.
Any opinions?
depends on what you want I guess.
if you want realtime, you should disable it.
as recommended in http://lac.linuxaudio.org/2010/recordin … ations.ogv
< Daenyth> and he works prolifically
4 8 15 16 23 42
Offline