Intel Atom D510MO performance issue
Hi, I was attempting to test temperatures under load by running "cat /dev/urandom > file" and I thought my system had crashed. Instantly when this command begins the system becomes very unresponsive. All input over ssh and keyboard attached to the machine has seemingly varying, but significant, amounts of lag. I'm not sure what would be causing this behaviour or how to properly diagnose it. The system in question is using the Intel Atom D510MO motherboard ( http://www.intel.com/products/desktop/motherboards/D510MO/D510MO-overview.htm ). I'm not sure how relevant this is, but the top output seems to indicate something: load averages: 1.73, 0.97, 0.46 03:18:14 33 processes: 30 idle, 3 on processor CPU0 states: 0.0% user, 0.0% nice, 0.4% system, 99.6% interrupt, 0.0% idle CPU1 states: 0.0% user, 0.0% nice, 6.3% system, 53.6% interrupt, 40.1% idle CPU2 states: 0.0% user, 0.0% nice, 100% system, 0.0% interrupt, 0.0% idle CPU3 states: 0.0% user, 0.0% nice, 3.7% system, 50.5% interrupt, 45.8% idle Memory: Real: 10M/137M act/tot Free: 845M Swap: 0K/2051M used/tot PID USERNAME PRI NICE SIZE RES STATE WAIT TIMECPU COMMAND 16216 root -50 184K 148K sleep/2 getblk 11:18 1129.88% cat 15668 _ntp 560 708K 812K onproc/0 - 0:42 77.64% ntpd 2808 root 20 624K 724K sleep/3 poll 0:17 60.40% ntpd 6584 root 320 808K 1392K onproc/1 - 0:08 9.23% top 16545 root 320 968K 1420K onproc/3 - 0:04 8.15% sendmail 23931 will 20 3340K 1812K sleep/1 select0:01 2.54% sshd 197 _syslogd 20 488K 644K sleep/1 poll 0:00 1.76% syslogd 2449 _pflogd40 540K 292K sleep/3 bpf 0:01 1.71% pflogd 17554 will 20 3200K 1804K idle select0:03 0.83% sshd This is after leaving the process running for a few minutes. top updates maybe once every minute while this goes on and the % cpu time slowly increases I'm aware this is a newish piece of equipment and may not be fully supported yet so I'm not sure if that is the reason or there is some bug here. Or that this is anything that should be too worried about. I initially found this behaviour on 4.6-release but then tried the Feb 23? (or which is on ftp as of a couple hours ago) snapshot. Same behaviour on both. Both were i386. I haven't yet tried amd64. Also, I tried a different test running infinite loops to max out all the CPUs and the system seemed to behave fine. Oh, I just tried the SP kernel while writing this and the problem only seems to occur when running the MP kernel. Sorry for the noise if I'm missing something. Here's the dmesg: OpenBSD 4.7-beta (GENERIC.MP) #423: Tue Feb 23 12:24:22 MST 2010 dera...@i386.openbsd.org:/usr/src/sys/arch/i386/compile/GENERIC.MP RTC BIOS diagnostic error 80 cpu0: Intel(R) Atom(TM) CPU D510 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 GHz cpu0: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CX16,xTPR real mem = 1055203328 (1006MB) avail mem = 1013702656 (966MB) RTC BIOS diagnostic error 80 mainbus0 at root bios0 at mainbus0: AT/286+ BIOS, date 11/17/09, SMBIOS rev. 2.5 @ 0xe4410 (25 entries) bios0: vendor Intel Corp. version "MOPNV10J.86A.0154.2009.1117.1624" date 11/17/2009 bios0: Intel Corporation D510MO acpi0 at bios0: rev 2 acpi0: tables DSDT FACP APIC MCFG HPET SSDT acpi0: wakeup devices SLPB(S4) PS2M(S4) PS2K(S4) UAR1(S4) UAR2(S4) P32_(S4) ILAN(S4) PEX0(S4) PEX1(S4) PEX2(S4) PEX3(S4) UHC1(S3) UHC2(S3) UHC3(S3) UHC4(S3) EHCI(S3) AZAL(S4) acpitimer0 at acpi0: 3579545 Hz, 24 bits acpimadt0 at acpi0 addr 0xfee0: PC-AT compat cpu0 at mainbus0: apid 0 (boot processor) cpu0: apic clock running at 166MHz cpu1 at mainbus0: apid 1 (application processor) cpu1: Intel(R) Atom(TM) CPU D510 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 GHz cpu1: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CX16,xTPR cpu2 at mainbus0: apid 2 (application processor) cpu2: Intel(R) Atom(TM) CPU D510 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 GHz cpu2: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CX16,xTPR cpu3 at mainbus0: apid 3 (application processor) cpu3: Intel(R) Atom(TM) CPU D510 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 GHz cpu3: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CX16,xTPR ioapic0 at mainbus0: apid 8 pa 0xfec0, version 20, 24 pins ioapic0: misconfigured as apic 0, remapped to apid 8 acpihpet0 at acpi0: 14318179 Hz acpiprt0 at acpi0: bus 5 (P32_) acpiprt1 at acpi0: bus 0 (PCI0) acpiprt2 at acpi0: bus 1 (PEX0) acpiprt3 at acpi0: bus 2 (PEX1) acpiprt4 at acpi0: bus 3 (PEX2) acpiprt5 at acpi0
Re: Intel Atom D510MO performance issue
Post output of 'vmstat -i' and read this thread http://marc.info/?l=openbsd-misc&m=126203835608528&w=2 On Thu, Feb 25, 2010 at 12:47 PM, Will Storey wrote: > Hi, > > I was attempting to test temperatures under load by running "cat > /dev/urandom > file" and I thought my system had crashed. Instantly when > this command begins the system becomes very unresponsive. All input over ssh > and keyboard attached to the machine has seemingly varying, but significant, > amounts of lag. > > I'm not sure what would be causing this behaviour or how to properly > diagnose it. > The system in question is using the Intel Atom D510MO motherboard ( > http://www.intel.com/products/desktop/motherboards/D510MO/D510MO-overview.htm > ). > > I'm not sure how relevant this is, but the top output seems to indicate > something: > > load averages: B 1.73, B 0.97, B 0.46 > 03:18:14 > 33 processes: B 30 idle, 3 on processor > CPU0 states: B 0.0% user, B 0.0% nice, B 0.4% system, 99.6% interrupt, B 0.0% > idle > CPU1 states: B 0.0% user, B 0.0% nice, B 6.3% system, 53.6% interrupt, 40.1% > idle > CPU2 states: B 0.0% user, B 0.0% nice, B 100% system, B 0.0% interrupt, B 0.0% > idle > CPU3 states: B 0.0% user, B 0.0% nice, B 3.7% system, 50.5% interrupt, 45.8% > idle > Memory: Real: 10M/137M act/tot B Free: 845M B Swap: 0K/2051M used/tot > > B PID USERNAME PRI NICE B SIZE B RES STATE B B WAIT B B B TIME B B CPU COMMAND > 16216 root B B B -5 B B 0 B 184K B 148K sleep/2 B getblk B 11:18 1129.88% cat > 15668 _ntp B B B 56 B B 0 B 708K B 812K onproc/0 B - B B B B 0:42 77.64% ntpd > B 2808 root B B B 2 B B 0 B 624K B 724K sleep/3 B poll B B B 0:17 60.40% ntpd > B 6584 root B B B 32 B B 0 B 808K 1392K onproc/1 B - B B B B 0:08 B 9.23% top > 16545 root B B B 32 B B 0 B 968K 1420K onproc/3 B - B B B B 0:04 B 8.15% sendmail > 23931 will B B B 2 B B 0 3340K 1812K sleep/1 B select B B 0:01 B 2.54% sshd > B 197 _syslogd B 2 B B 0 B 488K B 644K sleep/1 B poll B B B 0:00 B 1.76% syslogd > B 2449 _pflogd B B 4 B B 0 B 540K B 292K sleep/3 B bpf B B B 0:01 B 1.71% pflogd > 17554 will B B B 2 B B 0 3200K 1804K idle B B B select B B 0:03 B 0.83% sshd > > This is after leaving the process running for a few minutes. top updates > maybe once every minute while this goes on and the % cpu time slowly > increases > > I'm aware this is a newish piece of equipment and may not be fully supported > yet so I'm not sure if that is the reason or there is some bug here. Or that > this is anything that should be too worried about. > > I initially found this behaviour on 4.6-release but then tried the Feb 23? > (or which is on ftp as of a couple hours ago) snapshot. Same behaviour on > both. Both were i386. I haven't yet tried amd64. > > Also, I tried a different test running infinite loops to max out all the > CPUs and the system seemed to behave fine. > > Oh, I just tried the SP kernel while writing this and the problem only seems > to occur when running the MP kernel. > > Sorry for the noise if I'm missing something. > > Here's the dmesg: > > OpenBSD 4.7-beta (GENERIC.MP) #423: Tue Feb 23 12:24:22 MST 2010 > B B dera...@i386.openbsd.org:/usr/src/sys/arch/i386/compile/GENERIC.MP > RTC BIOS diagnostic error 80 > cpu0: Intel(R) Atom(TM) CPU D510 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 > GHz > cpu0: > FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUS H,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CX16,xTPR > real mem B = 1055203328 (1006MB) > avail mem = 1013702656 (966MB) > RTC BIOS diagnostic error 80 > mainbus0 at root > bios0 at mainbus0: AT/286+ BIOS, date 11/17/09, SMBIOS rev. 2.5 @ 0xe4410 > (25 entries) > bios0: vendor Intel Corp. version "MOPNV10J.86A.0154.2009.1117.1624" date > 11/17/2009 > bios0: Intel Corporation D510MO > acpi0 at bios0: rev 2 > acpi0: tables DSDT FACP APIC MCFG HPET SSDT > acpi0: wakeup devices SLPB(S4) PS2M(S4) PS2K(S4) UAR1(S4) UAR2(S4) P32_(S4) > ILAN(S4) PEX0(S4) PEX1(S4) PEX2(S4) PEX3(S4) UHC1(S3) UHC2(S3) UHC3(S3) > UHC4(S3) EHCI(S3) AZAL(S4) > acpitimer0 at acpi0: 3579545 Hz, 24 bits > acpimadt0 at acpi0 addr 0xfee0: PC-AT compat > cpu0 at mainbus0: apid 0 (boot processor) > cpu0: apic clock running at 166MHz > cpu1 at mainbus0: apid 1 (application processor) > cpu1: Intel(R) Atom(TM) CPU D510 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 > GHz > cpu1: > FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUS H,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CX16,xTPR > cpu2 at mainbus0: apid 2 (application processor) > cpu2: Intel(R) Atom(TM) CPU D510 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 > GHz > cpu2: > FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUS H,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CX16,xTPR > cpu3 at mainbus0: apid 3 (application processor) > cpu3: Intel(R) Atom(TM) CPU D510 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 > GH
Re: Intel Atom D510MO performance issue
On Thu, 25 Feb 2010 03:47:24 -0800 Will Storey wrote: > I'm not sure how relevant this is, but the top output seems to indicate > something: > > Memory: Real: 10M/137M act/tot Free: 845M Swap: 0K/2051M used/tot > > PID USERNAME PRI NICE SIZE RES STATE WAIT TIMECPU COMMAND > 16216 root -50 184K 148K sleep/2 getblk 11:18 1129.88% cat > 15668 _ntp 560 708K 812K onproc/0 - 0:42 77.64% ntpd > 2808 root 20 624K 724K sleep/3 poll 0:17 60.40% ntpd 1129.88%, 77.64%, and 60.40% CPU usage?
Re: Intel Atom D510MO performance issue
On 02/25/2010 06:47 AM, Will Storey wrote: Hi, I was attempting to test temperatures under load by running "cat /dev/urandom> file" and I thought my system had crashed. Instantly when this command begins the system becomes very unresponsive. All input over ssh and keyboard attached to the machine has seemingly varying, but significant, amounts of lag. I'm not sure what would be causing this behaviour or how to properly diagnose it. The system in question is using the Intel Atom D510MO motherboard ( http://www.intel.com/products/desktop/motherboards/D510MO/D510MO-overview.htm ). I'm not sure how relevant this is, but the top output seems to indicate something: load averages: 1.73, 0.97, 0.46 03:18:14 33 processes: 30 idle, 3 on processor CPU0 states: 0.0% user, 0.0% nice, 0.4% system, 99.6% interrupt, 0.0% idle CPU1 states: 0.0% user, 0.0% nice, 6.3% system, 53.6% interrupt, 40.1% idle CPU2 states: 0.0% user, 0.0% nice, 100% system, 0.0% interrupt, 0.0% idle CPU3 states: 0.0% user, 0.0% nice, 3.7% system, 50.5% interrupt, 45.8% idle Memory: Real: 10M/137M act/tot Free: 845M Swap: 0K/2051M used/tot PID USERNAME PRI NICE SIZE RES STATE WAIT TIMECPU COMMAND 16216 root -50 184K 148K sleep/2 getblk 11:18 1129.88% cat 15668 _ntp 560 708K 812K onproc/0 - 0:42 77.64% ntpd 2808 root 20 624K 724K sleep/3 poll 0:17 60.40% ntpd 6584 root 320 808K 1392K onproc/1 - 0:08 9.23% top 16545 root 320 968K 1420K onproc/3 - 0:04 8.15% sendmail 23931 will 20 3340K 1812K sleep/1 select0:01 2.54% sshd 197 _syslogd 20 488K 644K sleep/1 poll 0:00 1.76% syslogd 2449 _pflogd40 540K 292K sleep/3 bpf 0:01 1.71% pflogd 17554 will 20 3200K 1804K idle select0:03 0.83% sshd This is after leaving the process running for a few minutes. top updates maybe once every minute while this goes on and the % cpu time slowly increases I'm aware this is a newish piece of equipment and may not be fully supported yet so I'm not sure if that is the reason or there is some bug here. Or that this is anything that should be too worried about. I initially found this behaviour on 4.6-release but then tried the Feb 23? (or which is on ftp as of a couple hours ago) snapshot. Same behaviour on both. Both were i386. I haven't yet tried amd64. Also, I tried a different test running infinite loops to max out all the CPUs and the system seemed to behave fine. Oh, I just tried the SP kernel while writing this and the problem only seems to occur when running the MP kernel. Sorry for the noise if I'm missing something. Here's the dmesg: OpenBSD 4.7-beta (GENERIC.MP) #423: Tue Feb 23 12:24:22 MST 2010 dera...@i386.openbsd.org:/usr/src/sys/arch/i386/compile/GENERIC.MP RTC BIOS diagnostic error 80 cpu0: Intel(R) Atom(TM) CPU D510 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 GHz cpu0: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CX16,xTPR real mem = 1055203328 (1006MB) avail mem = 1013702656 (966MB) RTC BIOS diagnostic error 80 mainbus0 at root bios0 at mainbus0: AT/286+ BIOS, date 11/17/09, SMBIOS rev. 2.5 @ 0xe4410 (25 entries) bios0: vendor Intel Corp. version "MOPNV10J.86A.0154.2009.1117.1624" date 11/17/2009 bios0: Intel Corporation D510MO acpi0 at bios0: rev 2 acpi0: tables DSDT FACP APIC MCFG HPET SSDT acpi0: wakeup devices SLPB(S4) PS2M(S4) PS2K(S4) UAR1(S4) UAR2(S4) P32_(S4) ILAN(S4) PEX0(S4) PEX1(S4) PEX2(S4) PEX3(S4) UHC1(S3) UHC2(S3) UHC3(S3) UHC4(S3) EHCI(S3) AZAL(S4) acpitimer0 at acpi0: 3579545 Hz, 24 bits acpimadt0 at acpi0 addr 0xfee0: PC-AT compat cpu0 at mainbus0: apid 0 (boot processor) cpu0: apic clock running at 166MHz cpu1 at mainbus0: apid 1 (application processor) cpu1: Intel(R) Atom(TM) CPU D510 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 GHz cpu1: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CX16,xTPR cpu2 at mainbus0: apid 2 (application processor) cpu2: Intel(R) Atom(TM) CPU D510 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 GHz cpu2: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CX16,xTPR cpu3 at mainbus0: apid 3 (application processor) cpu3: Intel(R) Atom(TM) CPU D510 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 GHz cpu3: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CX16,xTPR ioapic0 at mainbus0: apid 8 pa 0xfec0, version 20, 24 pins ioapic0: misconfigured as apic 0, remapped to apid 8 acpihpet0 at acpi0: 14318179 Hz acpiprt0 at acpi0: bus 5 (P32_) acpiprt1 at acpi0: bus 0 (PCI0) acpiprt2 at acpi0: bus 1 (PEX0) acpiprt3 at acpi0: bus 2 (PEX1
Re: Intel Atom D510MO performance issue
Use arandom. On Thu, Feb 25, 2010 at 6:47 AM, Will Storey wrote: > Hi, > > I was attempting to test temperatures under load by running "cat > /dev/urandom > file" and I thought my system had crashed. Instantly when > this command begins the system becomes very unresponsive. All input over ssh > and keyboard attached to the machine has seemingly varying, but significant, > amounts of lag.