[Asterisk-Users] te110p and interrupts
Guys. I have an issue with a te110p card and also some tdm04b cards on the same system: Zttest returns this for the tdm04b cards: [EMAIL PROTECTED] ~]# /usr/src/zaptel-1.2.4/zttest 38 -v Opened pseudo zap interface, measuring accuracy... 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8223 sample intervals 99.621582% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8400 sample intervals 97.460938% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8438 sample intervals 96.997070% --- Results after 15 passes --- Best: 100.00 -- Worst: 96.997070 -- Average: 99.601237 And for the te110p: [EMAIL PROTECTED] ~]# /usr/src/zaptel-1.2.4/zttest 1 -v Opened pseudo zap interface, measuring accuracy... 8192 samples in 8239 sample intervals 99.426270% 8192 samples in 8200 sample intervals 99.902344% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8384 sample intervals 97.656250% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8193 sample intervals 99.987793% --- Results after 20 passes --- Best: 100.00 -- Worst: 97.656250 -- Average: 99.843140 Im getting problems receiving faxes which can be caused by this. As you can see, the cards are not sharing IRQs as far as I cant tell CPU0 CPU1 0: 17697848 17714488IO-APIC-edge timer 7: 2 2IO-APIC-edge parport0 8: 1 0IO-APIC-edge rtc 9: 0 1 IO-APIC-level acpi 14: 103455 112606IO-APIC-edge ide0 15: 0 0IO-APIC-edge libata 169: 0 0 IO-APIC-level uhci_hcd:usb2 185:2793868 12 IO-APIC-level eth0 201: 0 0 IO-APIC-level ehci_hcd:usb1 209: 0 0 IO-APIC-level uhci_hcd:usb3 217: 19672684 15531101 IO-APIC-level wcte11xp 225: 17824822 17371238 IO-APIC-level wctdm 233: 15323598 19896224 IO-APIC-level wctdm So, can anybody point me in the right direction in order to debug the problem with zttest and quality of sound? How can I change prioritiess for the cards or maybe adjust or see if they get their right number of interrupts, etc? Thx for the help guys! ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] te110p and interrupts
use lspci -vb for detecting interrupt conflicts..On 4/10/06, Anton Krall <[EMAIL PROTECTED] > wrote:Guys. I have an issue with a te110p card and also some tdm04b cards on the same system:Zttest returns this for the tdm04b cards:[EMAIL PROTECTED] ~]# /usr/src/zaptel-1.2.4/zttest 38 -vOpened pseudo zap interface, measuring accuracy...8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8192 sample intervals 100.00%8192 samples in 8223 sample intervals 99.621582%8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8192 sample intervals 100.00%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8400 sample intervals 97.460938%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8192 sample intervals 100.00%8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8438 sample intervals 96.997070%--- Results after 15 passes ---Best: 100.00 -- Worst: 96.997070 -- Average: 99.601237And for the te110p:[EMAIL PROTECTED] ~]# /usr/src/zaptel-1.2.4/zttest 1 -v Opened pseudo zap interface, measuring accuracy...8192 samples in 8239 sample intervals 99.426270%8192 samples in 8200 sample intervals 99.902344%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8384 sample intervals 97.656250%8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8191 sample intervals 99.987793%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8192 sample intervals 100.00%8192 samples in 8192 sample intervals 100.00%8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8193 sample intervals 99.987793%--- Results after 20 passes ---Best: 100.00 -- Worst: 97.656250 -- Average: 99.843140Im getting problems receiving faxes which can be caused by this.As you can see, the cards are not sharing IRQs as far as I cant tell CPU0 CPU1 0: 17697848 17714488IO-APIC-edge timer 7: 2 2IO-APIC-edge parport0 8: 1 0IO-APIC-edge rtc 9: 0 1 IO-APIC-level acpi 14: 103455 112606IO-APIC-edge ide0 15: 0 0IO-APIC-edge libata169: 0 0 IO-APIC-level uhci_hcd:usb2185:2793868 12 IO-APIC-level eth0201: 0 0 IO-APIC-level ehci_hcd:usb1209: 0 0 IO-APIC-level uhci_hcd:usb3217: 19672684 15531101 IO-APIC-level wcte11xp225: 17824822 17371238 IO-APIC-level wctdm233: 15323598 19896224 IO-APIC-level wctdmSo, can anybody point me in the right direction in order to debug the problem with zttest and quality of sound?How can I change prioritiess for the cards or maybe adjust or see if theyget their right number of interrupts, etc?Thx for the help guys!___ --Bandwidth and Colocation provided by Easynews.com --Asterisk-Users mailing listTo UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
RE: [Asterisk-Users] te110p and interrupts
This system has acpi enabled. This is how the cards looks like: 217: 22201132 17764597 IO-APIC-level wcte11xp225: 20253291 19707036 IO-APIC-level wctdm233: 17510466 22474179 IO-APIC-level wctdm but if you run lspci .vb to see how the system sees the cards: 02:02.0 Network controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface Subsystem: Unknown device 795e:0001 Flags: bus master, medium devsel, latency 32, IRQ 5 I/O ports at 2000 Memory at dd20 (32-bit, non-prefetchable) Capabilities: [40] Power Management version 2 02:08.0 Network controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface Subsystem: Unknown device b119:0001 Flags: bus master, medium devsel, latency 32, IRQ 5 I/O ports at 2800 Memory at dd201000 (32-bit, non-prefetchable) Capabilities: [40] Power Management version 2 03:02.0 Network controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface Subsystem: Unknown device b119:0001 Flags: bus master, medium devsel, latency 32, IRQ 5 I/O ports at 3000 Memory at dd40 (32-bit, non-prefetchable) Capabilities: [40] Power Management version 2 I dont get it, the system sees everything on IRQ5 eventhough the kernel sees everything in diff. IRQ: 02:02.0 Network controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface02:08.0 Network controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface03:02.0 Network controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface Also, latency for the cards is 32... i enabled the watchdog on zttool and I do see Missed IRQ/interrupts on zttooll for the te110p card... how can I fix this? Ive read some stuff around but so far.. nothing makes sense to me... :( From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Infobox PeruSent: Monday, April 10, 2006 5:47 PMTo: Asterisk Users Mailing List - Non-Commercial DiscussionSubject: Re: [Asterisk-Users] te110p and interrupts use lspci -vb for detecting interrupt conflicts.. On 4/10/06, Anton Krall <[EMAIL PROTECTED] > wrote: Guys. I have an issue with a te110p card and also some tdm04b cards on the same system:Zttest returns this for the tdm04b cards:[EMAIL PROTECTED] ~]# /usr/src/zaptel-1.2.4/zttest 38 -vOpened pseudo zap interface, measuring accuracy...8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8192 sample intervals 100.00%8192 samples in 8223 sample intervals 99.621582%8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8192 sample intervals 100.00%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8400 sample intervals 97.460938%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8192 sample intervals 100.00%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8438 sample intervals 96.997070%--- Results after 15 passes ---Best: 100.00 -- Worst: 96.997070 -- Average: 99.601237And for the te110p:[EMAIL PROTECTED] ~]# /usr/src/zaptel-1.2.4/zttest 1 -v Opened pseudo zap interface, measuring accuracy...8192 samples in 8239 sample intervals 99.426270%8192 samples in 8200 sample intervals 99.902344%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8384 sample intervals 97.656250%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8192 sample intervals 100.00%8192 samples in 8192 sample intervals 100.00%8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793%8192 samples in 8192 sample intervals 100.00%8192 samples in 8193 sample intervals 99.987793%--- Results after 20 passes ---Best: 100.00 -- Worst: 97.656250 -- Average: 99.843140Im getting problems receiving faxes which can be caused by this.As you can see, the cards are not sharing IRQs as far as I cant tell CPU0 CPU1 0: 17697848 17714488IO-APIC-edge timer 7: 2 2IO-APIC-edge parport0 8: 1 0IO-APIC-edge rtc 9: 0 1 I
Re: [Asterisk-Users] te110p and interrupts
Try booting with apic off, I think it's noapic kernel option. Notice this is "APIC" and not "ACPI", which you referred to. Then get your boards on different REAL irqs. Moj Anton Krall wrote: This system has acpi enabled. This is how the cards looks like: 217: 22201132 17764597 IO-APIC-level wcte11xp 225: 20253291 19707036 IO-APIC-level wctdm 233: 17510466 22474179 IO-APIC-level wctdm but if you run lspci .vb to see how the system sees the cards: 02:02.0 Network controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface Subsystem: Unknown device 795e:0001 Flags: bus master, medium devsel, latency 32, IRQ 5 I/O ports at 2000 Memory at dd20 (32-bit, non-prefetchable) Capabilities: [40] Power Management version 2 02:08.0 Network controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface Subsystem: Unknown device b119:0001 Flags: bus master, medium devsel, latency 32, IRQ 5 I/O ports at 2800 Memory at dd201000 (32-bit, non-prefetchable) Capabilities: [40] Power Management version 2 03:02.0 Network controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface Subsystem: Unknown device b119:0001 Flags: bus master, medium devsel, latency 32, IRQ 5 I/O ports at 3000 Memory at dd40 (32-bit, non-prefetchable) Capabilities: [40] Power Management version 2 I dont get it, the system sees everything on IRQ5 eventhough the kernel sees everything in diff. IRQ: 02:02.0 Network controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface 02:08.0 Network controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface 03:02.0 Network controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface Also, latency for the cards is 32... i enabled the watchdog on zttool and I do see Missed IRQ/interrupts on zttooll for the te110p card... how can I fix this? Ive read some stuff around but so far.. nothing makes sense to me... :( *From:* [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] *On Behalf Of *Infobox Peru *Sent:* Monday, April 10, 2006 5:47 PM *To:* Asterisk Users Mailing List - Non-Commercial Discussion *Subject:* Re: [Asterisk-Users] te110p and interrupts use lspci -vb for detecting interrupt conflicts.. On 4/10/06, *Anton Krall* <[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>> wrote: Guys. I have an issue with a te110p card and also some tdm04b cards on the same system: Zttest returns this for the tdm04b cards: [EMAIL PROTECTED] ~]# /usr/src/zaptel-1.2.4/zttest 38 -v Opened pseudo zap interface, measuring accuracy... 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8223 sample intervals 99.621582% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8400 sample intervals 97.460938% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8438 sample intervals 96.997070% --- Results after 15 passes --- Best: 100.00 -- Worst: 96.997070 -- Average: 99.601237 And for the te110p: [EMAIL PROTECTED] ~]# /usr/src/zaptel-1.2.4/zttest 1 -v Opened pseudo zap interface, measuring accuracy... 8192 samples in 8239 sample intervals 99.426270% 8192 samples in 8200 sample intervals 99.902344% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8384 sample intervals 97.656250% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99.987793% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8192 sample intervals 100.00% 8192 samples in 8191 sample intervals 99
RE: [Asterisk-Users] te110p and interrupts
Is this dual CPU/Core or just P4 with HT enabled? If it is P4, I would recommend to disable HT. Try changing PCI slots for one of the cards (if you have spare PCI slots). > CPU0 CPU1 > 0: 17697848 17714488IO-APIC-edge timer ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
RE: [Asterisk-Users] te110p and interrupts
I will try that and see what happens... This server is a supermicro one.. Anybody else had issues like this on supermicro? Any hints on how to resolv them? If I remember correctly, supermicro bios does let you assign irq to certain pci ports right? Will that help? Also, is there a way/is it recommended to increase the latency for these devices? |-Original Message- |From: [EMAIL PROTECTED] |[mailto:[EMAIL PROTECTED] On Behalf Of |Mojo with Horan & Company, LLC |Sent: Monday, April 10, 2006 6:20 PM |To: Asterisk Users Mailing List - Non-Commercial Discussion |Subject: Re: [Asterisk-Users] te110p and interrupts | |Try booting with apic off, I think it's noapic kernel option. |Notice this is "APIC" and not "ACPI", which you referred to. |Then get your boards on different REAL irqs. | |Moj | |Anton Krall wrote: |> This system has acpi enabled. This is how the cards looks like: |> |> 217: 22201132 17764597 IO-APIC-level wcte11xp |> 225: 20253291 19707036 IO-APIC-level wctdm |> 233: 17510466 22474179 IO-APIC-level wctdm |> |> but if you run lspci .vb to see how the system sees the cards: |> |> 02:02.0 Network controller: Tiger Jet Network Inc. Tiger3XX |Modem/ISDN |> interface |> Subsystem: Unknown device 795e:0001 |> Flags: bus master, medium devsel, latency 32, IRQ 5 |> I/O ports at 2000 |> Memory at dd20 (32-bit, non-prefetchable) |> Capabilities: [40] Power Management version 2 |02:08.0 Network |> controller: Tiger Jet Network Inc. Tiger3XX Modem/ISDN interface |> Subsystem: Unknown device b119:0001 |> Flags: bus master, medium devsel, latency 32, IRQ 5 |> I/O ports at 2800 |> Memory at dd201000 (32-bit, non-prefetchable) |> Capabilities: [40] Power Management version 2 |> |> 03:02.0 Network controller: Tiger Jet Network Inc. Tiger3XX |Modem/ISDN |> interface |> Subsystem: Unknown device b119:0001 |> Flags: bus master, medium devsel, latency 32, IRQ 5 |> I/O ports at 3000 |> Memory at dd40 (32-bit, non-prefetchable) |> Capabilities: [40] Power Management version 2 |> |> I dont get it, the system sees everything on IRQ5 eventhough the |> kernel sees everything in diff. IRQ: |> |> 02:02.0 Network controller: Tiger Jet Network Inc. Tiger3XX |Modem/ISDN |> interface 02:08.0 Network controller: Tiger Jet Network Inc. |Tiger3XX |> Modem/ISDN interface 03:02.0 Network controller: Tiger Jet Network |> Inc. Tiger3XX Modem/ISDN interface Also, latency for the cards is |> 32... |> |> i enabled the watchdog on zttool and I do see Missed |IRQ/interrupts on |> zttooll for the te110p card... how can I fix this? |> |> Ive read some stuff around but so far.. nothing makes sense to me... |> :( |> |> |> |--- |- |> *From:* [EMAIL PROTECTED] |> [mailto:[EMAIL PROTECTED] *On Behalf Of |> *Infobox Peru |> *Sent:* Monday, April 10, 2006 5:47 PM |> *To:* Asterisk Users Mailing List - Non-Commercial Discussion |> *Subject:* Re: [Asterisk-Users] te110p and interrupts |> |> use lspci -vb for detecting interrupt conflicts.. |> |> On 4/10/06, *Anton Krall* <[EMAIL PROTECTED] |> <mailto:[EMAIL PROTECTED]>> wrote: |> |> Guys. I have an issue with a te110p card and also some tdm04b |> cards on the |> same system: |> |> Zttest returns this for the tdm04b cards: |> |> [EMAIL PROTECTED] ~]# /usr/src/zaptel-1.2.4/zttest 38 -v |> Opened pseudo zap interface, measuring accuracy... |> |> 8192 samples in 8191 sample intervals 99.987793% |> 8192 samples in 8192 sample intervals 100.00% |> 8192 samples in 8192 sample intervals 100.00% |> 8192 samples in 8223 sample intervals 99.621582% |> 8192 samples in 8192 sample intervals 100.00% |> 8192 samples in 8191 sample intervals 99.987793% |> 8192 samples in 8192 sample intervals 100.00% |> 8192 samples in 8192 sample intervals 100.00% |> 8192 samples in 8191 sample intervals 99.987793% |> 8192 samples in 8400 sample intervals 97.460938% |> 8192 samples in 8191 sample intervals 99.987793% |> 8192 samples in 8192 sample intervals 100.00% |> 8192 samples in 8192 sample intervals 100.00% |> 8192 samples in 8191 sample intervals 99.987793% |> 8192 samples in 8438 sample intervals 96.997070% |> --- Results after 15 passes --- |> Best: 100.00 -- Worst: 96.997070 -- Average: 99.601237 |> |>
RE: [Asterisk-Users] te110p and interrupts
This is a Xeon with HT... I remember I disabled HT on supermicros bios but then again, maybe somebody enabled it again.. Thats why its showing 2 cpus? Ill disable it too.. Thx for the hint.. BTW why did you commented the timer line? |-Original Message- |From: [EMAIL PROTECTED] |[mailto:[EMAIL PROTECTED] On Behalf Of |Boris Bakchiev |Sent: Monday, April 10, 2006 6:26 PM |To: Asterisk Users Mailing List - Non-Commercial Discussion |Subject: RE: [Asterisk-Users] te110p and interrupts | |Is this dual CPU/Core or just P4 with HT enabled? |If it is P4, I would recommend to disable HT. | |Try changing PCI slots for one of the cards (if you have spare |PCI slots). | |> CPU0 CPU1 |> 0: 17697848 17714488IO-APIC-edge timer |___ |--Bandwidth and Colocation provided by Easynews.com -- | |Asterisk-Users mailing list |To UNSUBSCRIBE or update options visit: | http://lists.digium.com/mailman/listinfo/asterisk-users | | ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] te110p and interrupts
Because it was what showed you had two CPUs: CPU0, CPU1 (Hyperthreading is kinda dual-coreish) Anton Krall wrote: This is a Xeon with HT... I remember I disabled HT on supermicros bios but then again, maybe somebody enabled it again.. That’s why its showing 2 cpus? Ill disable it too.. Thx for the hint.. BTW why did you commented the timer line? |-Original Message- |From: [EMAIL PROTECTED] |[mailto:[EMAIL PROTECTED] On Behalf Of |Boris Bakchiev |Sent: Monday, April 10, 2006 6:26 PM |To: Asterisk Users Mailing List - Non-Commercial Discussion |Subject: RE: [Asterisk-Users] te110p and interrupts | |Is this dual CPU/Core or just P4 with HT enabled? |If it is P4, I would recommend to disable HT. | |Try changing PCI slots for one of the cards (if you have spare |PCI slots). | |> CPU0 CPU1 |> 0: 17697848 17714488IO-APIC-edge timer |___ |--Bandwidth and Colocation provided by Easynews.com -- | |Asterisk-Users mailing list |To UNSUBSCRIBE or update options visit: | http://lists.digium.com/mailman/listinfo/asterisk-users | | ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users -- Mojo <[EMAIL PROTECTED]> Office Manger, Horan & Company, LLC (907) 747- x112 ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] te110p and interrupts
Anton Krall wrote: I will try that and see what happens... This server is a supermicro one.. Anybody else had issues like this on supermicro? Any hints on how to resolv them? If I remember correctly, supermicro bios does let you assign irq to certain pci ports right? Will that help? Also, is there a way/is it recommended to increase the latency for these devices? We tried to use a supermicro board once. We could not make it work with Asterisk and a Digium TE card if there was any usage of the IDE controller. We made the system a new file server and went with an Intel branded board. ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
RE: [Asterisk-Users] te110p and interrupts
Damn :) any others? |-Original Message- |From: [EMAIL PROTECTED] |[mailto:[EMAIL PROTECTED] On Behalf Of |Eric "ManxPower" Wieling |Sent: Monday, April 10, 2006 7:25 PM |To: Asterisk Users Mailing List - Non-Commercial Discussion |Subject: Re: [Asterisk-Users] te110p and interrupts | |Anton Krall wrote: |> I will try that and see what happens... |> This server is a supermicro one.. Anybody else had issues |like this on |> supermicro? Any hints on how to resolv them? |> |> If I remember correctly, supermicro bios does let you assign irq to |> certain pci ports right? Will that help? |> |> Also, is there a way/is it recommended to increase the latency for |> these devices? | |We tried to use a supermicro board once. We could not make it |work with Asterisk and a Digium TE card if there was any usage |of the IDE controller. We made the system a new file server |and went with an Intel branded board. |___ |--Bandwidth and Colocation provided by Easynews.com -- | |Asterisk-Users mailing list |To UNSUBSCRIBE or update options visit: | http://lists.digium.com/mailman/listinfo/asterisk-users | | ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] te110p and interrupts
On Monday 10 April 2006 19:20, Mojo with Horan & Company, LLC wrote: > Try booting with apic off, I think it's noapic kernel option. Notice > this is "APIC" and not "ACPI", which you referred to. Then get your > boards on different REAL irqs. Please do not open your mouth to spout nonsense if you do not know what you're talking about. APIC interrupts are far more "real" than emulated XT-PIC interrupts. If the IO-APIC can put each device on its own "high" interrupt it means that that INT# signal on the PCI slot *is* on a totally separate, physical IRQ line which is routed into that APIC. When you disable native APIC mode you force it into compatibility mode, where it essentially performs a logical "OR" on the real, separate IRQ lines and gives you a single, edge-triggered i8259-style "low" IRQ. (In reality it's not a logical OR since the XT (i8259-style) interrupts are edge-triggered, not level triggered, but that's neither here nor there for this particular discussion.) Again, if the IO-APIC is reporting that the card is on its own IRQ, it really, truly, honestly *IS* on its own IRQ. The reason that it is suggested to disable the IO-APIC is that on many low-end systems, the IO-APIC is plain old broken and causing other issues. I don't think I've run across a system board in the last year or two with that issue, though. It's always been on older P3 and early P4 systems. Anton, your problem is very likely simple interrupt load. You have three Digium cards in there, and they're all generating their own 1000Hz interrupt. If you did the newbie thing and compiled your kernel with a HZ value of '1000' because you felt it would be better, you have that overhead as well. Your system is very likely just having trouble coping with so many interrupts. My personal opinion is that you should sell the three Digium cards and buy a single dualspan card and a cheap channel bank. Your interrupt load will drop by 2/3 and your system will be FAR happier. So check the kernel HZ value first; I set *all* my Asterisk systems to the old style HZ of 100; there is simply no need for anything more on a server, *especially* if you've already got hardware providing a real 1000Hz interrupt instead of ztdummy trying to emulate such a thing. -A. ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] te110p and interrupts
On Mon, 2006-04-10 at 19:25 -0500, Eric "ManxPower" Wieling wrote: > Anton Krall wrote: > > I will try that and see what happens... > > This server is a supermicro one.. Anybody else had issues like this on > > supermicro? Any hints on how to resolv them? > > > > If I remember correctly, supermicro bios does let you assign irq to certain > > pci ports right? Will that help? > > > > Also, is there a way/is it recommended to increase the latency for these > > devices? > > We tried to use a supermicro board once. We could not make it work with > Asterisk and a Digium TE card if there was any usage of the IDE > controller. We made the system a new file server and went with an Intel > branded board. I'm running a Supermicro system with a te110p and a tdm40b. I have HT turned off in the bios and I have "acpi=off noapic" added as boot options in grub. The cards work just fine. FYI, Kenny signature.asc Description: This is a digitally signed message part ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
RE: [Asterisk-Users] te110p and interrupts
Hi Andrew... Thank you very much for the info. I didn't recompile the kernel, Im using a generic 2.6 kernel but its worth taking a look at what you said.. Where can I find (which file) the Hz the kernel was precompiled to? Also, Im running 1 te110p and 2 tdm cards, probably I'll disable 1 card later but I will need at least 1 tdm and the te110p for my E1. So you suggest not disabling any apic/acpi stuff then.. How about HT? should I disable that on the bios? Im using supermicro servers and eventhough I get voice calls without problems, Im getting a few IRQ Misses from time to time which makes faxing on the E1 very difficult (the original problem). What do you think? |-Original Message- |From: [EMAIL PROTECTED] |[mailto:[EMAIL PROTECTED] On Behalf Of |Andrew Kohlsmith |Sent: Tuesday, April 11, 2006 7:37 AM |To: asterisk-users@lists.digium.com |Subject: Re: [Asterisk-Users] te110p and interrupts | |On Monday 10 April 2006 19:20, Mojo with Horan & Company, LLC wrote: |> Try booting with apic off, I think it's noapic kernel |option. Notice |> this is "APIC" and not "ACPI", which you referred to. Then get your |> boards on different REAL irqs. | |Please do not open your mouth to spout nonsense if you do not |know what you're talking about. | |APIC interrupts are far more "real" than emulated XT-PIC |interrupts. If the IO-APIC can put each device on its own |"high" interrupt it means that that INT# signal on the PCI |slot *is* on a totally separate, physical IRQ line which is |routed into that APIC. When you disable native APIC mode you |force it into compatibility mode, where it essentially |performs a logical "OR" on the real, separate IRQ lines and |gives you a single, edge-triggered i8259-style "low" IRQ. | |(In reality it's not a logical OR since the XT (i8259-style) |interrupts are edge-triggered, not level triggered, but that's |neither here nor there for this particular discussion.) | |Again, if the IO-APIC is reporting that the card is on its own |IRQ, it really, truly, honestly *IS* on its own IRQ. The |reason that it is suggested to disable the IO-APIC is that on |many low-end systems, the IO-APIC is plain old broken and |causing other issues. I don't think I've run across a system |board in the last year or two with that issue, though. It's |always been on older P3 and early P4 systems. | |Anton, your problem is very likely simple interrupt load. You |have three Digium cards in there, and they're all generating |their own 1000Hz interrupt. |If you did the newbie thing and compiled your kernel with a HZ |value of '1000' because you felt it would be better, you have |that overhead as well. | |Your system is very likely just having trouble coping with so |many interrupts. |My personal opinion is that you should sell the three Digium |cards and buy a single dualspan card and a cheap channel bank. | Your interrupt load will drop by 2/3 and your system will be |FAR happier. | |So check the kernel HZ value first; I set *all* my Asterisk |systems to the old style HZ of 100; there is simply no need |for anything more on a server, |*especially* if you've already got hardware providing a real |1000Hz interrupt instead of ztdummy trying to emulate such a thing. | |-A. |___ |--Bandwidth and Colocation provided by Easynews.com -- | |Asterisk-Users mailing list |To UNSUBSCRIBE or update options visit: | http://lists.digium.com/mailman/listinfo/asterisk-users | | ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
RE: [Asterisk-Users] te110p and interrupts
Zttool shows no irqmisses on the te110p card? |-Original Message- |From: [EMAIL PROTECTED] |[mailto:[EMAIL PROTECTED] On Behalf Of |Kenneth Lussier |Sent: Tuesday, April 11, 2006 7:05 AM |To: Asterisk Users Mailing List - Non-Commercial Discussion |Subject: Re: [Asterisk-Users] te110p and interrupts | |On Mon, 2006-04-10 at 19:25 -0500, Eric "ManxPower" Wieling wrote: |> Anton Krall wrote: |> > I will try that and see what happens... |> > This server is a supermicro one.. Anybody else had issues |like this |> > on supermicro? Any hints on how to resolv them? |> > |> > If I remember correctly, supermicro bios does let you |assign irq to |> > certain pci ports right? Will that help? |> > |> > Also, is there a way/is it recommended to increase the latency for |> > these devices? |> |> We tried to use a supermicro board once. We could not make it work |> with Asterisk and a Digium TE card if there was any usage of the IDE |> controller. We made the system a new file server and went with an |> Intel branded board. | |I'm running a Supermicro system with a te110p and a tdm40b. I |have HT turned off in the bios and I have "acpi=off noapic" |added as boot options in grub. The cards work just fine. | |FYI, |Kenny | | ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
R: [Asterisk-Users] te110p and interrupts
Hi Anton, I'm using a supermicro P4 3GHz P8SCT (Intel E7221 chipset) with TE205P and a TDM04 and I've similar problem. I was using linux 2.6.9smp that seams to have problem with APCI so Hyperthreading, even if enabled, was not working (I sow 1 cpu). Today I've disable hypertrading and start using mono processor kernel, disabled both serial and parallel interfaces (via bios setup), prevent loading of kernel module for second gigabit interface card and zaptel tdm module (I not using tdm04 interface at the moment). These bios changes also seams to have disabled APIC (is it related to hyperthreading ?). However, after all this I've a 4 busy hours of no "chan_zap.c: PRI got event: HDLC Bad FCS (8) on Primary D-channel of span 2" error. I'll make you know for news. Take also a look at http://www.voip-info.org/wiki/view/Asterisk+hardware , chrisre seams to had our same problem with TE110 and supermicro. Simone -Messaggio originale- Da: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Per conto di Anton Krall Inviato: martedì 11 aprile 2006 5.28 A: 'Asterisk Users Mailing List - Non-Commercial Discussion' Oggetto: RE: [Asterisk-Users] te110p and interrupts Damn :) any others? |-Original Message- |From: [EMAIL PROTECTED] |[mailto:[EMAIL PROTECTED] On Behalf Of Eric |"ManxPower" Wieling |Sent: Monday, April 10, 2006 7:25 PM |To: Asterisk Users Mailing List - Non-Commercial Discussion |Subject: Re: [Asterisk-Users] te110p and interrupts | |Anton Krall wrote: |> I will try that and see what happens... |> This server is a supermicro one.. Anybody else had issues |like this on |> supermicro? Any hints on how to resolv them? |> |> If I remember correctly, supermicro bios does let you assign irq to |> certain pci ports right? Will that help? |> |> Also, is there a way/is it recommended to increase the latency for |> these devices? | |We tried to use a supermicro board once. We could not make it work |with Asterisk and a Digium TE card if there was any usage of the IDE |controller. We made the system a new file server and went with an |Intel branded board. |___ |--Bandwidth and Colocation provided by Easynews.com -- | |Asterisk-Users mailing list |To UNSUBSCRIBE or update options visit: | http://lists.digium.com/mailman/listinfo/asterisk-users | | ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] te110p and interrupts
On Tue, 11 Apr 2006, Andrew Kohlsmith wrote: > Please do not open your mouth to spout nonsense if you do not know > what you're talking about. [...] > Again, if the IO-APIC is reporting that the card is on its own IRQ, > it really, truly, honestly *IS* on its own IRQ. The reason that it > is suggested to disable the IO-APIC is that on many low-end systems, > the IO-APIC is plain old broken and causing other issues. I don't > think I've run across a system board in the last year or two with > that issue, though. It's always been on older P3 and early P4 > systems. Allow me to comment that Digium actually recommends turning off APIC and using "lspci -vb" to troubleshoot this kind of shared-interrupt problem. Cheers, Gerald. ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] te110p and interrupts
Hi, I've been battling with a similar issue: a) I wrote a script to periodically run the command "cat /proc/interrupts" and figure out the interrupts per second. I run this script for over 24 hours and never once did the difference between the preceeding and succeeding interrupt counts go below 1005 (wierd result because of (b) below); b) zttest was reporting very bad results; c) lspci -vb was reporting that the TE110P shared an IRQ with the Gigabit Ethernet Card (IRQ 11) d) lspci -vv was reporting that the TE110P was on an IRQ of its own ("IRQ 24") probably because of APIC (wierd because of (c) above); e) Users reported intermittent bad audio; Below are the [experimental] steps I took: a) I'm running a Dual 3.2 GHz machine - the network card is services by CPU0 - I set the smp_affinity value for the Digium card to be CPU1 b) I disabled the userland 'irqbalance' process which keeps switching the Digium card between the CPUs c) I increased the PCI "LATENCY_TIMER" value for the TE110P to a value higher than the Gigabit Card. So far, things are looking quite good - zttest is reporting very encouraging worst-case figures when run over a period of over an hour (it reports 99.98% worst case at off peak time and 99.77% when run during the busy hour). Ultimately when I have physical access to the machine, I will change the PCI slots to see if getting "lspci -vb" to report that the card is on its own IRQ will improve performance further. Cheers, Gerald. ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] te110p and interrupts
On Tuesday 11 April 2006 21:45, Begumisa Gerald M wrote: > Hi, > > I've been battling with a similar issue: > > a) I wrote a script to periodically run the command "cat > /proc/interrupts" and figure out the interrupts per second. I run this > script for over 24 hours and never once did the difference between the > preceeding and succeeding interrupt counts go below 1005 (wierd result > because of (b) below); > > b) zttest was reporting very bad results; > > c) lspci -vb was reporting that the TE110P shared an IRQ with the Gigabit > Ethernet Card (IRQ 11) > > d) lspci -vv was reporting that the TE110P was on an IRQ of its own ("IRQ > 24") probably because of APIC (wierd because of (c) above); > > e) Users reported intermittent bad audio; > > > Below are the [experimental] steps I took: > > a) I'm running a Dual 3.2 GHz machine - the network card is services by > CPU0 - I set the smp_affinity value for the Digium card to be CPU1 I am curious.. Have you tried disabling CPU1 by setting isolcpus=1 on the kernel command line ? This will make the kernel ignore the second CPU - you can then run asterisk on it by using the taskset command (from schedutils) taskset 0x0001 asterisk -p and asterisk wlll run on a CPU all on its own. I was about to try this and wondered if you might give it a try and report back. > > b) I disabled the userland 'irqbalance' process which keeps switching the > Digium card between the CPUs > > c) I increased the PCI "LATENCY_TIMER" value for the TE110P to a value > higher than the Gigabit Card. > > > So far, things are looking quite good - zttest is reporting very > encouraging worst-case figures when run over a period of over an hour (it > reports 99.98% worst case at off peak time and 99.77% when run during the > busy hour). Ultimately when I have physical access to the machine, I will > change the PCI slots to see if getting "lspci -vb" to report that the card > is on its own IRQ will improve performance further. There is a lot of rubbish said about interrupts - if possible use the APIC - it is the new beast that solves interrupt starvation problems. Seeing if the card has an IRQ 5 as reported by lspci is misleading - this is the IRQ that would be used if operating in old mode and in old mode you will have to disable all kinds of things to make it work. In APIC mode, new hardware comes into play and each device can have its own interrupt - it is the way to go. The only problem is that uniprocessor APIC implementations are known to be badly done - recent motherboards should be OK. For elucidation see http://www.microsoft.com/whdc/system/sysperf/IO-APIC.mspx http://www.microsoft.com/whdc/system/sysperf/apic.mspx http://en.wikipedia.org/wiki/Advanced_Programmable_Interrupt_Controller http://en.wikipedia.org/wiki/Intel_APIC_Architecture Paul Hewlett ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] te110p and interrupts
Hi Paul, Thanks for the message! On Sun, 16 Apr 2006, Paul Hewlett wrote: [...] > I am curious.. > > Have you tried disabling CPU1 by setting isolcpus=1 on the kernel > command line ? > > This will make the kernel ignore the second CPU - you can then run > asterisk on it by using the taskset command (from schedutils) > > taskset 0x0001 asterisk -p > > and asterisk wlll run on a CPU all on its own. I was about to try > this and wondered if you might give it a try and report back. I haven't done this yet. Once we have physical access to the machine, I'll make sure we try this out and see what difference it makes. Cheers! Gerald. ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
RE: [Asterisk-Users] te110p and interrupts
I don't know if this only works with multiple cpus but I have HT enabled and it shows cpu0 and cpu1 .. I tried the first part of this email and still the kernel boots and shows 2 cpus.. Will this only work with 2 real cpus? |-Original Message- |From: [EMAIL PROTECTED] |[mailto:[EMAIL PROTECTED] On Behalf Of |Begumisa Gerald M |Sent: Monday, April 17, 2006 5:13 AM |To: Asterisk Users Mailing List - Non-Commercial Discussion |Subject: Re: [Asterisk-Users] te110p and interrupts | |Hi Paul, | |Thanks for the message! | | On Sun, 16 Apr 2006, Paul Hewlett wrote: | [...] |> I am curious.. |> |> Have you tried disabling CPU1 by setting isolcpus=1 on |the kernel |> command line ? |> |> This will make the kernel ignore the second CPU - you |can then run |> asterisk on it by using the taskset command (from schedutils) |> |> taskset 0x0001 asterisk -p |> |> and asterisk wlll run on a CPU all on its own. I was about to try |> this and wondered if you might give it a try and report back. | |I haven't done this yet. Once we have physical access to the |machine, I'll make sure we try this out and see what |difference it makes. | | |Cheers! |Gerald. |___ |--Bandwidth and Colocation provided by Easynews.com -- | |Asterisk-Users mailing list |To UNSUBSCRIBE or update options visit: | http://lists.digium.com/mailman/listinfo/asterisk-users | | ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] te110p and interrupts
On 4/11/06, Begumisa Gerald M <[EMAIL PROTECTED]> wrote: > > Again, if the IO-APIC is reporting that the card is on its own IRQ, > > it really, truly, honestly *IS* on its own IRQ. The reason that it > > is suggested to disable the IO-APIC is that on many low-end systems, > Allow me to comment that Digium actually recommends turning off APIC and > using "lspci -vb" to troubleshoot this kind of shared-interrupt problem. Interesting. Now 'why' do they suggest it, is it because older IO-APIC are 'broken' on some boards? I'm very curious as to 'why', because that would give everyone a better idea on what to look for when having this kind of problems. cheers ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] te110p and interrupts
On Mon, 17 Apr 2006, stoffell wrote: > Interesting. Now 'why' do they suggest it, is it because older > IO-APIC are 'broken' on some boards? I'm very curious as to 'why', [...] Most likely this is why. Regards, Gerald ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] te110p and interrupts
On Monday 17 April 2006 12:39, Anton Krall wrote: > I don't know if this only works with multiple cpus but I have HT enabled > and it shows cpu0 and cpu1 .. I tried the first part of this email and > still the kernel boots and shows 2 cpus.. Will this only work with 2 real > cpus? I believe so. Hyperthreading is not really SMP. Additionally I think that HT should not be used if asterisk is doing a lot of transcoding - whilst HT gives u 2 'pipes' there is still only one FPU and heavy transcoding will simply bottleneck at the FPU(NB SSE/MMX count as FP since they use the FP registers.). If u turn HT on, a 1% penalty is incurred fir the mutexes required so heavy use of the FPU will actually show a decrease in speed compared to no HT. In theory at least. :-) Paul Hewlett > > |-Original Message- > |From: [EMAIL PROTECTED] > |[mailto:[EMAIL PROTECTED] On Behalf Of > |Begumisa Gerald M > |Sent: Monday, April 17, 2006 5:13 AM > |To: Asterisk Users Mailing List - Non-Commercial Discussion > |Subject: Re: [Asterisk-Users] te110p and interrupts > | > |Hi Paul, > | > |Thanks for the message! > | > | On Sun, 16 Apr 2006, Paul Hewlett wrote: > | [...] > | > |> I am curious.. > |> > |> Have you tried disabling CPU1 by setting isolcpus=1 on > | > |the kernel > | > |> command line ? > |> > |> This will make the kernel ignore the second CPU - you > | > |can then run > | > |> asterisk on it by using the taskset command (from schedutils) > |> > |> taskset 0x0001 asterisk -p > |> > |> and asterisk wlll run on a CPU all on its own. I was about to try > |> this and wondered if you might give it a try and report back. > | > |I haven't done this yet. Once we have physical access to the > |machine, I'll make sure we try this out and see what > |difference it makes. > | > | > |Cheers! > |Gerald. > |___ > |--Bandwidth and Colocation provided by Easynews.com -- > | > |Asterisk-Users mailing list > |To UNSUBSCRIBE or update options visit: > | http://lists.digium.com/mailman/listinfo/asterisk-users > > ___ > --Bandwidth and Colocation provided by Easynews.com -- > > Asterisk-Users mailing list > To UNSUBSCRIBE or update options visit: >http://lists.digium.com/mailman/listinfo/asterisk-users ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
RE: [Asterisk-Users] te110p and interrupts
What you say makes a lot of sense... Well.. I disabled HT and Im still having irqmisses :( I stil have to try the latency idea.. Hope that helps.. |-Original Message- |From: [EMAIL PROTECTED] |[mailto:[EMAIL PROTECTED] On Behalf Of |Paul Hewlett |Sent: Monday, April 17, 2006 12:19 PM |To: asterisk-users@lists.digium.com |Subject: Re: [Asterisk-Users] te110p and interrupts | |On Monday 17 April 2006 12:39, Anton Krall wrote: |> I don't know if this only works with multiple cpus but I have HT |> enabled and it shows cpu0 and cpu1 .. I tried the first part of this |> email and still the kernel boots and shows 2 cpus.. Will this only |> work with 2 real cpus? | | I believe so. Hyperthreading is not really SMP. | | Additionally I think that HT should not be used if asterisk |is doing a lot of transcoding - whilst HT gives u 2 'pipes' |there is still only one FPU and heavy transcoding will simply |bottleneck at the FPU(NB SSE/MMX count as FP since they use |the FP registers.). If u turn HT on, a 1% penalty is incurred |fir the mutexes required so heavy use of the FPU will actually |show a decrease in speed compared to no HT. In theory at least. :-) | |Paul Hewlett |> |> |-Original Message- |> |From: [EMAIL PROTECTED] |> |[mailto:[EMAIL PROTECTED] On Behalf Of |> |Begumisa Gerald M |> |Sent: Monday, April 17, 2006 5:13 AM |> |To: Asterisk Users Mailing List - Non-Commercial Discussion |> |Subject: Re: [Asterisk-Users] te110p and interrupts |> | |> |Hi Paul, |> | |> |Thanks for the message! |> | |> | On Sun, 16 Apr 2006, Paul Hewlett wrote: |> | [...] |> | |> |> I am curious.. |> |> |> |> Have you tried disabling CPU1 by setting isolcpus=1 on |> | |> |the kernel |> | |> |> command line ? |> |> |> |> This will make the kernel ignore the second CPU - you |> | |> |can then run |> | |> |> asterisk on it by using the taskset command (from schedutils) |> |> |> |> taskset 0x0001 asterisk -p |> |> |> |> and asterisk wlll run on a CPU all on its own. I was |about to try |> |> this and wondered if you might give it a try and report back. |> | |> |I haven't done this yet. Once we have physical access to |the machine, |> |I'll make sure we try this out and see what difference it makes. |> | |> | |> |Cheers! |> |Gerald. ___ --Bandwidth and Colocation provided by Easynews.com -- Asterisk-Users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users