64bit kills SMP ? - Gentoo Forums

64bit kills SMP ?

Locked

36 posts

1
2
Next

Author

Message

acoul: n00b; Posts: 38; Joined: Wed Jan 19, 2005 8:13 am; Contact:
Contact acoul

Website

64bit kills SMP ?

Quote

Post by acoul » Sat Dec 08, 2007 11:19 am

I have three AMD64 X2 systems and noticed that the processes do not get evenly distributed among the two CPUs. This is true for the 64bit systems. The system that is on a 32bit setup distributes quite evenly the processes among the two CPUs. The configuration for the 64bit system is based on the AMD64 X2 wiki. Here is the output of the /proc/interrupts of each system (the last one is the 32bit system):

Code: Select all

Linux version 2.6.23-gentoo-r3 (root@orama) (gcc version 4.1.2 (Gentoo 4.1.2 p1.0.2)) #1 SMP Mon Dec 3 18:21:41 EET 2007

vendor_id       : AuthenticAMD
cpu family      : 15
model           : 67
model name      : AMD Athlon(tm) 64 X2 Dual Core Processor 6000+
stepping        : 3
cpu MHz         : 3000.000
cache size      : 1024 KB

           CPU0       CPU1
  0:       8728  103367201   IO-APIC-edge      timer
  1:          2       3005   IO-APIC-edge      i8042
  6:          0          3   IO-APIC-edge      floppy
  8:          0          1   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 12:          0        148   IO-APIC-edge      i8042
 16:          0          1   IO-APIC-fasteoi   ohci_hcd:usb2
 17:          0          2   IO-APIC-fasteoi   ohci_hcd:usb3, ohci_hcd:usb5
 18:          0          2   IO-APIC-fasteoi   ohci_hcd:usb4, ohci_hcd:usb6
 19:        576    2431625   IO-APIC-fasteoi   ehci_hcd:usb1, eth0
 22:        152     545508   IO-APIC-fasteoi   ahci
NMI:          0          0
LOC:  103378656  103378633
ERR:          0

Code: Select all

Linux version 2.6.23-gentoo-r3 (root@lexy) (gcc version 4.1.2 (Gentoo 4.1.2)) #2 SMP PREEMPT Thu Dec 6 09:40:49 EET 2007

vendor_id       : AuthenticAMD
cpu family      : 15
model           : 72
model name      : AMD Turion(tm) 64 X2 Mobile Technology TL-50
stepping        : 2
cpu MHz         : 1600.000
cache size      : 256 KB

           CPU0       CPU1
  0:      37023   15790384   IO-APIC-edge      timer
  1:        180      58446   IO-APIC-edge      i8042
  9:        569     195499   IO-APIC-fasteoi   acpi
 12:       9198    2771145   IO-APIC-edge      i8042
 14:          0         38   IO-APIC-edge      ide0
 18:       6987    3879570   IO-APIC-fasteoi   nvidia
 19:          0          0   IO-APIC-fasteoi   ohci_hcd:usb2
 20:        293      35595   IO-APIC-fasteoi   HDA Intel
 21:          0          0   IO-APIC-fasteoi   ehci_hcd:usb1
 22:       3036     708270   IO-APIC-fasteoi   sata_nv
 23:      27318   11838332   IO-APIC-fasteoi   eth0
NMI:          0          0
LOC:   15828128   15828102
ERR:          0

Code: Select all

Linux version 2.6.23-gentoo-r3 (root@aifnis) (gcc version 4.1.2 (Gentoo 4.1.2 p1.0.1)) #1 SMP Tue Dec 4 10:55:50 EET 2007

vendor_id       : AuthenticAMD
cpu family      : 15
model           : 75
model name      : AMD Athlon(tm) 64 X2 Dual Core Processor 3800+
stepping        : 2
cpu MHz         : 2034.377
cache size      : 512 KB

           CPU0       CPU1
  0:         84          0   IO-APIC-edge      timer
  1:          0          8   IO-APIC-edge      i8042
  8:          0          0   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 12:        100          4   IO-APIC-edge      i8042
 14:    1417823    1128208   IO-APIC-edge      ide0
 16:   53907210   58287007   IO-APIC-fasteoi   eth1
 17:          0          0   IO-APIC-fasteoi   sata_nv
 18:     795062     249175   IO-APIC-fasteoi   ide2, ide3
 19:          0          0   IO-APIC-fasteoi   ohci_hcd:usb1
 20:          0          0   IO-APIC-fasteoi   ehci_hcd:usb2
NMI:          0          0
LOC:   87781973   87781975
ERR:          1
MIS:          0

is this a known issue or am I missing something?

Udasai: n00b; Posts: 32; Joined: Mon Sep 20, 2004 5:47 pm

Quote

Post by Udasai » Sat Dec 08, 2007 5:14 pm

I have a similar distribution. It makes sense if CPU0 is being used more for regular threads and the "spare" CPU1 gets assigned the interrupts. It also makes sense from a caching perspective to keep a recurrent interrupt on a single processor. I'm just speculating here, though.

loftwyr: l33t; Posts: 970; Joined: Wed Dec 29, 2004 2:51 am; Location: 43°38'23.62"N 79°27'8.60"W

Quote

Post by loftwyr » Sat Dec 08, 2007 6:25 pm

You need to install irqbalance. The x86-64 kernel code doesn't balance IRQs natively so they moved it to a userspace tool.

My emerge --info
Have you run revdep-rebuild lately? It's in gentoolkit and it's worth a shot if things don't work well.
Celebrating 5 years of Gentoo-ing.

AaronPPC: Guru; Posts: 522; Joined: Sun May 29, 2005 8:22 am; Location: Tucson, AZ

Quote

Post by AaronPPC » Sat Dec 08, 2007 8:52 pm

Forgive the stupid question, but what is the advantage of balancing IRQs across CPUs?

--Aaron

loftwyr: l33t; Posts: 970; Joined: Wed Dec 29, 2004 2:51 am; Location: 43°38'23.62"N 79°27'8.60"W

Quote

Post by loftwyr » Sat Dec 08, 2007 10:26 pm

For most people, very little but if you're running CPU intensive apps, having balanced usage smooths things out.

My emerge --info
Have you run revdep-rebuild lately? It's in gentoolkit and it's worth a shot if things don't work well.
Celebrating 5 years of Gentoo-ing.

cyrre: n00b; Posts: 26; Joined: Mon Jul 02, 2007 8:01 am

Quote

Post by cyrre » Sat Dec 08, 2007 11:32 pm

Seems like it's working fine for me.

The CPU is a Intel Quad Q6600 running in 64bit.

Code: Select all

cat /proc/interrupts 

           CPU0       CPU1       CPU2       CPU3       
  0:     113703     113825     113678     113783   IO-APIC-edge      timer
  1:          0          0          0          2   IO-APIC-edge      i8042
  6:          0          2          2          1   IO-APIC-edge      floppy
  8:          1          0          0          0   IO-APIC-edge      rtc
  9:          0          0          0          0   IO-APIC-fasteoi   acpi
 14:      11991      11911      11897      11964   IO-APIC-edge      libata
 15:        511        513        521        505   IO-APIC-edge      libata
 16:      30746      30638      31051      30621   IO-APIC-fasteoi   libata, uhci_hcd:usb3, eth0, nvidia
 18:       4670       4625       4702       4620   IO-APIC-fasteoi   ide3, ehci_hcd:usb1, uhci_hcd:usb5, uhci_hcd:usb8
 19:          0          0          0          0   IO-APIC-fasteoi   ahci, libata, uhci_hcd:usb7
 21:       5896       5925       5682       5753   IO-APIC-fasteoi   uhci_hcd:usb4
 22:      19175      19140      19048      19331   IO-APIC-fasteoi   HDA Intel
 23:          0          2          0          0   IO-APIC-fasteoi   ehci_hcd:usb2, uhci_hcd:usb6
NMI:          0          0          0          0 
LOC:     453094     452568     453042     452513 
ERR:          0

streamkid: Tux's lil' helper; Posts: 82; Joined: Tue Jul 17, 2007 6:53 am; Location: Greece

Quote

Post by streamkid » Sun Dec 09, 2007 8:23 am

I should say it's pretty ok for me too. Maybe there is something with your installation?

Code: Select all

beast streamkid # cat /proc/interrupts 
           CPU0       CPU1       
  0:    4794587    4792755   IO-APIC-edge      timer
  1:      11819      11811   IO-APIC-edge      i8042
  8:          1          0   IO-APIC-edge      rtc
  9:      62538      62596   IO-APIC-fasteoi   acpi
 12:         61         62   IO-APIC-edge      i8042
 14:     172415     172450   IO-APIC-edge      ide0
 16:    2122370    2123687   IO-APIC-fasteoi   uhci_hcd:usb3, ipw3945, i915@pci:0000:00:02.0
 17:     120713     120870   IO-APIC-fasteoi   eth0
 18:          1          1   IO-APIC-fasteoi   ehci_hcd:usb1, uhci_hcd:usb7
 19:       2079       1982   IO-APIC-fasteoi   uhci_hcd:usb6
 20:          1          1   IO-APIC-fasteoi   ohci1394
 21:     116075     115977   IO-APIC-fasteoi   uhci_hcd:usb4, sdhci:slot0
 22:        855       1022   IO-APIC-fasteoi   HDA Intel
 23:         11         12   IO-APIC-fasteoi   ehci_hcd:usb2, uhci_hcd:usb5
1276:      61840      62080   PCI-MSI-edge      ahci
NMI:          0          0 
LOC:    9587304    9587280 
ERR:          0

Intel C2D 7100

LoSeR_5150: Guru; Posts: 455; Joined: Sun Mar 20, 2005 9:19 pm; Location: San Francisco, CA

Quote

Post by LoSeR_5150 » Tue Dec 11, 2007 12:04 am

On my Core2Duo T7200 (32bit install) and Core2Duo 6600 (64bit install) both were not balancing irq's correctly or at least not like streamkid or cyrre. After installing irqbalance and starting it (etc/init.d/irqbalance start) they seem to balancing much better. When I get home I'll check my amd machines as I expect to see the same behavior. After balancing the irqs the system does feel 'snappier' but it may just be placebo effect

Opteron 1356@2.4Ghz
6GB DDR2 800Mhz
128MB Quadro NVS 210S
640GB Western Digital HD
*Gentoo-x86_64-2.6.30-r1

Opteron175@2.2GHz
2GB DDR 400MHz
256MB Quadro 1400 Go
(2) 80GB Segate HDs: RAID0
*Gentoo-x86_64-2.6.30-r1

LoSeR_5150: Guru; Posts: 455; Joined: Sun Mar 20, 2005 9:19 pm; Location: San Francisco, CA

Quote

Post by LoSeR_5150 » Tue Dec 11, 2007 7:57 pm

I forgot that my 32bit install is only a single core

but on my opteron the irq balances is WAY different then my core2duo at work.

Opteron w/o irqbalance

Code: Select all

loser@Hell ~ $ cat /proc/interrupts 
           CPU0       CPU1       
  0:   23293439   12866466   IO-APIC-edge      timer
  1:        953       9305   IO-APIC-edge      i8042
  6:          0          2   IO-APIC-edge      floppy
  7:          0          0   IO-APIC-edge      parport0
  8:          0          0   IO-APIC-edge      rtc
 10:      14169       6447   IO-APIC-fasteoi   acpi
 12:       2109       4125   IO-APIC-edge      i8042
 15:     209357     115111   IO-APIC-edge      ide1
 17:     237618     175047   IO-APIC-fasteoi   wifi0
 18:          0          1   IO-APIC-fasteoi   yenta
 20:       6526      76642   IO-APIC-fasteoi   sata_via, ohci1394
 21:     765581     518221   IO-APIC-fasteoi   ehci_hcd:usb1, uhci_hcd:usb2, uhci_hcd:usb3, uhci_hcd:usb4, uhci_hcd:usb5
 22:          0          0   IO-APIC-fasteoi   VIA8237
 24:     321802     639436   IO-APIC-fasteoi   nvidia
NMI:          0          0 
LOC:   36155472   36154990 
ERR:          0

Core2Duo w/ irqbalance

Code: Select all

jeremy@Diesel ~ $ cat /proc/interrupts 
           CPU0       CPU1       
  0:   72610043          0   IO-APIC-edge      timer
  1:          2          0   IO-APIC-edge      i8042
  8:          1          0   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 12:          4          0   IO-APIC-edge      i8042
 14:     529146          0   IO-APIC-edge      ide0
 16:        812    4620575   IO-APIC-fasteoi   HDA Intel, nvidia
 18:        146      41357   IO-APIC-fasteoi   uhci_hcd:usb4
 21:         40     273852   IO-APIC-fasteoi   ehci_hcd:usb1, uhci_hcd:usb2
 22:      26809          0   IO-APIC-fasteoi   uhci_hcd:usb3
 23:          0          0   IO-APIC-fasteoi   uhci_hcd:usb5
314:        129   11498450   PCI-MSI-edge      eth0
315:       5525     157759   PCI-MSI-edge      ahci
NMI:          0          0 
LOC:   71640309   71640236 
ERR:          0

I just enabled irqbalance on the opteron, how long should I wait to see what the interrupts look like? Like do I need to wait for it to "settle" ? Thanks

Opteron 1356@2.4Ghz
6GB DDR2 800Mhz
128MB Quadro NVS 210S
640GB Western Digital HD
*Gentoo-x86_64-2.6.30-r1

Opteron175@2.2GHz
2GB DDR 400MHz
256MB Quadro 1400 Go
(2) 80GB Segate HDs: RAID0
*Gentoo-x86_64-2.6.30-r1

loftwyr: l33t; Posts: 970; Joined: Wed Dec 29, 2004 2:51 am; Location: 43°38'23.62"N 79°27'8.60"W

Quote

Post by loftwyr » Tue Dec 11, 2007 10:34 pm

IRQ balance doesn't just split the interrupts between processors evenly. It does it by groups, making sure devices that should be hitting an interrupt at different times are one different processes for added smoothness.

My emerge --info
Have you run revdep-rebuild lately? It's in gentoolkit and it's worth a shot if things don't work well.
Celebrating 5 years of Gentoo-ing.

Cyker: Veteran; Posts: 1746; Joined: Thu Jun 15, 2006 7:43 pm

Quote

Post by Cyker » Wed Dec 12, 2007 8:17 am

IIRC, irqbalance or not doesn't matter on a single-socket CPUs and has a minimal (if any) effect beyond making /proc/interrupts look nicer.

I could be mis-remembering 'tho.

Belliash: Advocate; Posts: 2503; Joined: Wed Nov 24, 2004 1:39 pm; Location: Wroclaw, Poland; Contact:
Contact Belliash

Website

Quote

Post by Belliash » Wed Dec 12, 2007 9:56 am

how to enable IRQ balancing?

Asio Software Technologies
Belliash IT Weblog

gringo: Advocate; Posts: 3793; Joined: Sun Apr 27, 2003 10:25 am

Quote

Post by gringo » Wed Dec 12, 2007 2:26 pm

Morpheouss wrote:how to enable IRQ balancing?

emerge -v irqbalance && rc-update add irqbalance boot.

cheers

Error: Failing not supported by current locale

acoul: n00b; Posts: 38; Joined: Wed Jan 19, 2005 8:13 am; Contact:
Contact acoul

Website

Quote

Post by acoul » Mon Dec 17, 2007 2:30 pm

loftwyr wrote:You need to install irqbalance. The x86-64 kernel code doesn't balance IRQs natively so they moved it to a userspace tool.

thank you.

acoul: n00b; Posts: 38; Joined: Wed Jan 19, 2005 8:13 am; Contact:
Contact acoul

Website

Quote

Post by acoul » Wed Dec 19, 2007 12:54 pm

even with the irqbalance application installed the irqs do not get evenly utilized by both CPUs. This to my understanding means that the two CPUs do not get fully utilized by the system.

Code: Select all

           CPU0       CPU1
  0:       2480    2659750   IO-APIC-edge      timer
  1:          0         63   IO-APIC-edge      i8042
  9:       3199       4564   IO-APIC-fasteoi   acpi
 12:          1        122   IO-APIC-edge      i8042
 14:          2         58   IO-APIC-edge      libata
 15:          0          0   IO-APIC-edge      libata
 18:     399193     176793   IO-APIC-fasteoi   nvidia
 19:       2738        160   IO-APIC-fasteoi   HDA Intel
 20:          0          4   IO-APIC-fasteoi   ehci_hcd:usb2
 21:      30237      67410   IO-APIC-fasteoi   ohci_hcd:usb1
 22:     514691       2318   IO-APIC-fasteoi   sata_nv
 23:          1    1164723   IO-APIC-fasteoi   eth0
NMI:          0          0
LOC:    2662316    2662293
ERR:          0

on a core2 duo, 32-bit system, the interrupts do get evenly utilized by both CPUs:

Code: Select all

vendor_id       : GenuineIntel
cpu family      : 6
model           : 15
model name      : Intel(R) Core(TM)2 CPU          4300  @ 1.80GHz
stepping        : 2
cpu MHz         : 1795.539
cache size      : 2048 KB

Code: Select all

Linux version 2.6.23-gentoo-r3 (root@orion) (gcc version 4.1.2 (Gentoo 4.1.2 p1.0.1)) #1 SMP Wed Dec 5 11:07:05 EET 2007

Code: Select all

           CPU0       CPU1
  0:         85          0   IO-APIC-edge      timer
  1:       3156          0   IO-APIC-edge      i8042
  8:          0          0   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 12:          4          0   IO-APIC-edge      i8042
 14:    1842872    2090628   IO-APIC-edge      ide0
 16:          0          0   IO-APIC-fasteoi   uhci_hcd:usb5
 17:   45988621   45950060   IO-APIC-fasteoi   eth1
 18:          0          0   IO-APIC-fasteoi   uhci_hcd:usb4
 19:     307278     438757   IO-APIC-fasteoi   libata, uhci_hcd:usb3
 20:          0          0   IO-APIC-fasteoi   ehci_hcd:usb1, uhci_hcd:usb2
 21:     383244     220286   IO-APIC-fasteoi   ide2
NMI:          0          0
LOC:   48840672   48840670
ERR:          0
MIS:          0

am I missing something?

loftwyr: l33t; Posts: 970; Joined: Wed Dec 29, 2004 2:51 am; Location: 43°38'23.62"N 79°27'8.60"W

Quote

Post by loftwyr » Wed Dec 19, 2007 2:12 pm

It's a long read but it explains everything
http://irqbalance.org/documentation.php

My emerge --info
Have you run revdep-rebuild lately? It's in gentoolkit and it's worth a shot if things don't work well.
Celebrating 5 years of Gentoo-ing.

acoul: n00b; Posts: 38; Joined: Wed Jan 19, 2005 8:13 am; Contact:
Contact acoul

Website

Quote

Post by acoul » Thu Dec 20, 2007 7:40 pm

a very good reading.

DualCpuUser: n00b; Posts: 71; Joined: Thu Mar 03, 2005 1:03 am

Quote

Post by DualCpuUser » Sat Jan 05, 2008 10:39 pm

For power mgmt you push things to 1 proc (core) so you can lower the C-States of the other, I to had seen this on my E6000 and Qx6700, but when you take into consideration of sleeping processors for as long a possible that makes sense, and also cache locality.

Bouncing processes and such between processors causes a lot cache trashing. I have read that link in the past, and as I type this , I do recall many of the issues involved.

I never got to play with irqbalance when my gentoo was up and running I had to dump it 9 months ago.

-Ron

-Ron

snIP3r: l33t; Posts: 866; Joined: Fri May 21, 2004 6:21 pm; Location: germany

Quote

Post by snIP3r » Sat Mar 01, 2008 8:21 pm

hi!

i also installed the irqbalance package. before installing irqbalance i got this output to /proc/interrupts:

Code: Select all

area52 conf.d # cat /proc/interrupts
           CPU0       CPU1
  0:      75379  481568417   IO-APIC-edge      timer
  1:          3       5789   IO-APIC-edge      i8042
  4:        723     596597   IO-APIC-edge      serial
  8:          0          1   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 16:      29168   17751056   IO-APIC-fasteoi   3w-9xxx, serial
 19:          0          3   IO-APIC-fasteoi   ohci1394
 20:       7473    4362242   IO-APIC-fasteoi   ohci_hcd:usb1
 21:          0          0   IO-APIC-fasteoi   sata_nv
 22:          0          0   IO-APIC-fasteoi   sata_nv
 23:          2         58   IO-APIC-fasteoi   sata_nv
1272:        237     193572   PCI-MSI-edge      eth1
1273:        117      96725   PCI-MSI-edge      eth0
NMI:          0          0
LOC:  481647773  481647747
ERR:          0

can anyone tell me how long it takes for irqbalance to adapt?

greets
snIP3r

Intel i3-4130T on ASUS P9D-X
Kernel 6.6.52-gentoo SMP
-----------------------------------------------
if your problem is fixed please add something like [solved] to the topic!

loftwyr: l33t; Posts: 970; Joined: Wed Dec 29, 2004 2:51 am; Location: 43°38'23.62"N 79°27'8.60"W

Quote

Post by loftwyr » Sat Mar 01, 2008 10:32 pm

irqbalance takes effect immediately. You should notice a significant difference very quickly.

My emerge --info
Have you run revdep-rebuild lately? It's in gentoolkit and it's worth a shot if things don't work well.
Celebrating 5 years of Gentoo-ing.

tallica: Apprentice; Posts: 152; Joined: Fri Jul 27, 2007 2:52 pm; Location: Lublin, POL

Quote

Post by tallica » Wed Mar 05, 2008 3:32 pm

I can't get it to work...

Code: Select all

$ cat /proc/interrupts 
           CPU0       CPU1       
  0:        105          0   IO-APIC-edge      timer
  1:       3032          0   IO-APIC-edge      i8042
  6:          5          0   IO-APIC-edge      floppy
  8:          1          0   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 16:     239625          0   IO-APIC-fasteoi   uhci_hcd:usb5, nvidia
 18:          0          0   IO-APIC-fasteoi   uhci_hcd:usb4
 19:          0          0   IO-APIC-fasteoi   uhci_hcd:usb3
 22:        777          0   IO-APIC-fasteoi   hdaudio0
 23:     690304          0   IO-APIC-fasteoi   ehci_hcd:usb1, uhci_hcd:usb2
315:      94625          0   PCI-MSI-edge      ahci
316:       6632          0   PCI-MSI-edge      eth0
NMI:          0          0   Non-maskable interrupts
LOC:     490722     355345   Local timer interrupts
RES:      31547      30161   Rescheduling interrupts
CAL:        102        203   function call interrupts
TLB:        946       1180   TLB shootdowns
TRM:          0          0   Thermal event interrupts
THR:          0          0   Threshold APIC interrupts
SPU:          0          0   Spurious interrupts
ERR:          0

Here is my kernel config: http://tallica.pl/linux/.config
dmesg: http://tallica.pl/linux/dmesg
emerge --info: http://tallica.pl/linux/emerge_info

Same with and without irqbalance running...

Gentoo ~AMD64 | Audacious

snIP3r: l33t; Posts: 866; Joined: Fri May 21, 2004 6:21 pm; Location: germany

Quote

Post by snIP3r » Wed Mar 05, 2008 3:59 pm

hi all!

here are my actual results:

Code: Select all

area52 ~ # cat /proc/interrupts
           CPU0       CPU1
  0:      90640  564171240   IO-APIC-edge      timer
  1:         23       6088   IO-APIC-edge      i8042
  4:        723     712620   IO-APIC-edge      serial
  8:          0          1   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 16:      29168   21847206   IO-APIC-fasteoi   3w-9xxx, serial
 19:          0          3   IO-APIC-fasteoi   ohci1394
 20:     791295    4362242   IO-APIC-fasteoi   ohci_hcd:usb1
 21:          0          0   IO-APIC-fasteoi   sata_nv
 22:          0          0   IO-APIC-fasteoi   sata_nv
 23:          2         58   IO-APIC-fasteoi   sata_nv
1272:        237     193572   PCI-MSI-edge      eth1
1273:        117      96725   PCI-MSI-edge      eth0
NMI:          0          0
LOC:  564266545  564266519
ERR:          0

comparing this with my previous post, it looks like it is working...

@Michal_86: your kernel config looks ok for me, the 2 cores are recognized by the kernel. i also wonder why your 2.nd core is absolutely not used... even without the irqbalance package installed, the 2nd core should get something to do... i assume its something in the kernel config. could you perhaps boot form a live-cd and check if its working there?

greets
snIP3r

Intel i3-4130T on ASUS P9D-X
Kernel 6.6.52-gentoo SMP
-----------------------------------------------
if your problem is fixed please add something like [solved] to the topic!

tallica: Apprentice; Posts: 152; Joined: Fri Jul 27, 2007 2:52 pm; Location: Lublin, POL

Quote

Post by tallica » Wed Mar 05, 2008 5:11 pm

Ubuntu LiveCD 7.10 AMD64

Code: Select all

Linux ubuntu 2.6.22-14-generic #1 SMP Sun Oct 14 21:45:15 GMT 2007 x86_64 GNU/Linux

$ cat /proc/interrupts 
           CPU0       CPU1       
  0:      75028          0   IO-APIC-edge      timer
  1:        115          0   IO-APIC-edge      i8042
  6:          9          0   IO-APIC-edge      floppy
  8:          0          0   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 12:          4          0   IO-APIC-edge      i8042
 16:          0          0   IO-APIC-fasteoi   uhci_hcd:usb4
 17:         49          0   IO-APIC-fasteoi   eth0
 18:          0          0   IO-APIC-fasteoi   uhci_hcd:usb3
 19:       8200          0   IO-APIC-fasteoi   uhci_hcd:usb2, ahci
 22:       1001          0   IO-APIC-fasteoi   HDA Intel
 23:       4128          0   IO-APIC-fasteoi   uhci_hcd:usb1, ehci_hcd:usb5
NMI:          0          0 
LOC:      74515      74497 
ERR:          0

And with irqbalance running:

Code: Select all

$ cat /proc/interrupts 
           CPU0       CPU1       
  0:     173059          0   IO-APIC-edge      timer
  1:        269        490   IO-APIC-edge      i8042
  6:          9          0   IO-APIC-edge      floppy
  8:          0          0   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 12:          4          0   IO-APIC-edge      i8042
 16:          0          0   IO-APIC-fasteoi   uhci_hcd:usb4
 17:       9754         31   IO-APIC-fasteoi   eth0
 18:          0          0   IO-APIC-fasteoi   uhci_hcd:usb3
 19:       9560        118   IO-APIC-fasteoi   uhci_hcd:usb2, ahci
 22:       1001          0   IO-APIC-fasteoi   HDA Intel
 23:      11199          0   IO-APIC-fasteoi   uhci_hcd:usb1, ehci_hcd:usb5
NMI:          0          0 
LOC:     172023     172005 
ERR:          0

Edit:
Ubuntu LiveCD 7.04 i686

Code: Select all

Linux ubuntu 2.6.20-15-generic #2 SMP Sun Apr 15 07:36:31 UTC 2007 i686 GNU/Linux

$ cat /proc/interrupts 
           CPU0       CPU1       
  0:      48822          0   IO-APIC-edge      timer
  1:         41          0   IO-APIC-edge      i8042
  6:          9          0   IO-APIC-edge      floppy
  8:         50          0   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 12:          4          0   IO-APIC-edge      i8042
 16:          0          0   IO-APIC-fasteoi   uhci_hcd:usb4
 17:       4955       2595   IO-APIC-fasteoi   eth0
 18:       3210          0   IO-APIC-fasteoi   uhci_hcd:usb1, ehci_hcd:usb5
 19:       8692          0   IO-APIC-fasteoi   uhci_hcd:usb2, libata
 20:          0          0   IO-APIC-fasteoi   uhci_hcd:usb3
 21:        803          0   IO-APIC-fasteoi   HDA Intel
NMI:          0          0 
LOC:      48471      48466 
ERR:          0
MIS:          0

irqbalance running:

Code: Select all

$ cat /proc/interrupts 
           CPU0       CPU1       
  0:     105782          0   IO-APIC-edge      timer
  1:        182        189   IO-APIC-edge      i8042
  6:          9          0   IO-APIC-edge      floppy
  8:         50          0   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 12:          4          0   IO-APIC-edge      i8042
 16:          0          0   IO-APIC-fasteoi   uhci_hcd:usb4
 17:       5002       2610   IO-APIC-fasteoi   eth0
 18:       5653          0   IO-APIC-fasteoi   uhci_hcd:usb1, ehci_hcd:usb5
 19:       9011         71   IO-APIC-fasteoi   uhci_hcd:usb2, libata
 20:          0          0   IO-APIC-fasteoi   uhci_hcd:usb3
 21:        803          0   IO-APIC-fasteoi   HDA Intel
NMI:          0          0 
LOC:     105115     105110 
ERR:          0
MIS:          0

Gentoo ~AMD64 | Audacious

darkphader: Veteran; Posts: 1225; Joined: Thu May 09, 2002 11:24 pm; Location: Motown; Contact:
Contact darkphader

Website

Quote

Post by darkphader » Wed Mar 05, 2008 7:55 pm

Decided to try irqbalance but it seems to only work in --oneshot fashion. It starts and then stops, no error messages are provided. Anyone else seeing this?

EDIT UPDATE: looks like it's because I don't have more than one socket - it's a dual-core system - irqbalance not desirable.

WYSIWYG - What You See Is What You Grep

loftwyr: l33t; Posts: 970; Joined: Wed Dec 29, 2004 2:51 am; Location: 43°38'23.62"N 79°27'8.60"W

Quote

Post by loftwyr » Wed Mar 05, 2008 10:50 pm

irqbalance is for dual core systems.

Can we see the relevant log entries?

My emerge --info
Have you run revdep-rebuild lately? It's in gentoolkit and it's worth a shot if things don't work well.
Celebrating 5 years of Gentoo-ing.

Locked

36 posts

1
2
Next

Return to “Gentoo on AMD64”