Sign in with
Sign up | Sign in
Your question
Solved

Random single-core spikes in AMD 8120 CPU

Tags:
  • Core
  • Processors
  • CPUs
  • AMD
Last response: in CPUs
Share
November 19, 2013 3:50:10 PM

Specs:
Motherboard: GA-970A-D3 Gigabyte
Processor: AMD FX(tm)-8120 Eight-core processor
RAM: 8GB (2x 4GB Viper Xtreme)
Ethernet: RTL8111/8168B PCI Express Gigabit
HDD: ST33000651AS Seagate 3TB
Power Supply: Antec 620
GPU: Asus Radeon HD 7770

So, I had this desktop custom built by a friend of mine a little while ago. I didn't experience any issues with it for the first handful of months, but at some point, without any hardware modification (heck, never even opened it to clean it up yet) I started experiencing some pretty weird CPU spikes.

http://postimg.org/image/dl66cay5t/

A single core will peak to nearly 100%, and then they'll alternate between eachother. Some will randomly go from ''parked'' to ''unparked'', and everything just generally turns into a mess.

The problem doesn't seem to behave in any predictable way. Sometimes it starts while I'm watching a movie with absolutely nothing else running, sometimes I'll be editing HD videos with no issues whatsoever.

But if I even attempt to play any video game that's slightly demanding, it starts doing its thing, making my CPU's temp rise to worrying, -but not quite dangerous- levels, and the CPU fan starts sounding like an upset jet engine.

I have tried a few things already; installing different OS (Win 7, XP, several versions of linux, currently running Kubuntu) wiping my HDD clean, messing around with drivers... But I can't put my finger on the source.

I almost don't want to believe it's a hardware problem, because I don't see why it would start out of nowhere like it did, with no changes done to my build at all, but on the software side of things; Nothing to be found.

Ask for any further specifications or details and I'll provide them, I've already done my share of googling, but since I'm not exactly experienced, and since this issue is so bizarre, I'm not even sure where to start.

More about : random single core spikes amd 8120 cpu

a c 473 à CPUs
a c 117 À AMD
November 19, 2013 4:30:07 PM

First thing you could try is a simple antivirus/malware scan.
Secondly, get a can of compressed air and clean the inside of that machine! If the fan's clogged with dust you will get high temperatures which will lead to the noise.
Thirdly, there are some patches for Windows 7 that resolve some issues for Bulldozer FX CPUs like your own, give me a sec to find them.

Try measuring your CPU temps with a program like HWmonitor or HWinfo and report back.
m
0
l
Related resources
November 19, 2013 4:44:00 PM

JOOK-D said:
First thing you could try is a simple antivirus/malware scan.
Secondly, get a can of compressed air and clean the inside of that machine! If the fan's clogged with dust you will get high temperatures which will lead to the noise.
Thirdly, there are some patches for Windows 7 that resolve some issues for Bulldozer FX CPUs like your own, give me a sec to find them.

Try measuring your CPU temps with a program like HWmonitor or HWinfo and report back.


I think you missed some lines in my post.
I wiped the drive clean, and tried several different OS, which didn't fix it at all, so that eliminates the malware issue (unless they can resist a complete disk formatting..?)
Also I'm not even running windows right now, as I said, so those patches are not for me.
I've also I cleaned my machine a handful of times since then, to no avail. And even if it was clogged, shouldn't the high CPU usage then be constant, and not such prominent random bumps?
I measured the temperatures with Speedfan (and anotherr software at some point), and at its peak it hovered around 69-71C. Usually somewhere around 36C when idle. My case is a well ventilated Antec Eleven hundred. There shouldn't be any issues ''normally''.

Edit: Here's another screenshot, taken just now, right after publishing this post; http://postimg.org/image/3kpv7x6yh/
m
0
l

Best solution

a c 473 à CPUs
a c 117 À AMD
November 19, 2013 4:55:54 PM

This is the reason I shouldn't be on the forums at 1am haha. My bad.

Eeeesh, well you eliminated most of the common causes, seems you have quite an isolated case here.

Another potential cause could be as simple as a usb device in a port, try moving them around. Other than that, make sure all of the cables are plugged into the motherboard firmly. Pretty generic advice, not sure what to say.

I'm pretty stumped. No strange sounds or anything?
Share
November 19, 2013 5:15:13 PM

JOOK-D said:

I'm pretty stumped. No strange sounds or anything?


Not really, no. I'll mess around with the cables a little more, altough I didn't see anything out of place last I checked. Perhaps I'll ask someone irl to check it out (said ''friend'' who built it for me would be a good start I assume)
Thanks for the input anyway!
If anyone else has a clue.. Anything is welcomed. I thought that core-specific bumps like those would be significative of a specific issue, but oh well... Here goes my wishful thinking for the day!
m
0
l
November 21, 2013 6:22:40 PM

I ran cpufrequtils, and here's the output. Not quite sure how to interpret that but it might be usefull info, maybe..?
CPU 6 is the only one running at 3.10 GHz, but I assume this is normal and has to do with the others being ''parked'' or something.


  1. analyzing CPU 0:
  2. driver: acpi-cpufreq
  3. CPUs which run at the same hardware frequency: 0
  4. CPUs which need to have their frequency coordinated by software: 0
  5. maximum transition latency: 4.0 us.
  6. hardware limits: 1.40 GHz - 3.10 GHz
  7. available frequency steps: 3.10 GHz, 2.80 GHz, 2.30 GHz, 1.90 GHz, 1.40 GHz
  8. available cpufreq governors: conservative, ondemand, userspace, powersave, performance
  9. current policy: frequency should be within 1.40 GHz and 3.10 GHz.
  10. The governor "ondemand" may decide which speed to use
  11. within this range.
  12. current CPU frequency is 1.40 GHz.
  13. cpufreq stats: 3.10 GHz:-nan%, 2.80 GHz:-nan%, 2.30 GHz:-nan%, 1.90 GHz:-nan%, 1.40 GHz:-nan% (95732)
  14. analyzing CPU 1:
  15. driver: acpi-cpufreq
  16. CPUs which run at the same hardware frequency: 1
  17. CPUs which need to have their frequency coordinated by software: 1
  18. maximum transition latency: 4.0 us.
  19. hardware limits: 1.40 GHz - 3.10 GHz
  20. available frequency steps: 3.10 GHz, 2.80 GHz, 2.30 GHz, 1.90 GHz, 1.40 GHz
  21. available cpufreq governors: conservative, ondemand, userspace, powersave, performance
  22. current policy: frequency should be within 1.40 GHz and 3.10 GHz.
  23. The governor "ondemand" may decide which speed to use
  24. within this range.
  25. current CPU frequency is 1.40 GHz.
  26. cpufreq stats: 3.10 GHz:-nan%, 2.80 GHz:-nan%, 2.30 GHz:-nan%, 1.90 GHz:-nan%, 1.40 GHz:-nan% (102437)
  27. analyzing CPU 2:
  28. driver: acpi-cpufreq
  29. CPUs which run at the same hardware frequency: 2
  30. CPUs which need to have their frequency coordinated by software: 2
  31. maximum transition latency: 4.0 us.
  32. hardware limits: 1.40 GHz - 3.10 GHz
  33. available frequency steps: 3.10 GHz, 2.80 GHz, 2.30 GHz, 1.90 GHz, 1.40 GHz
  34. available cpufreq governors: conservative, ondemand, userspace, powersave, performance
  35. current policy: frequency should be within 1.40 GHz and 3.10 GHz.
  36. The governor "ondemand" may decide which speed to use
  37. within this range.
  38. current CPU frequency is 1.40 GHz.
  39. cpufreq stats: 3.10 GHz:-nan%, 2.80 GHz:-nan%, 2.30 GHz:-nan%, 1.90 GHz:-nan%, 1.40 GHz:-nan% (69474)
  40. analyzing CPU 3:
  41. driver: acpi-cpufreq
  42. CPUs which run at the same hardware frequency: 3
  43. CPUs which need to have their frequency coordinated by software: 3
  44. maximum transition latency: 4.0 us.
  45. hardware limits: 1.40 GHz - 3.10 GHz
  46. available frequency steps: 3.10 GHz, 2.80 GHz, 2.30 GHz, 1.90 GHz, 1.40 GHz
  47. available cpufreq governors: conservative, ondemand, userspace, powersave, performance
  48. current policy: frequency should be within 1.40 GHz and 3.10 GHz.
  49. The governor "ondemand" may decide which speed to use
  50. within this range.
  51. current CPU frequency is 1.40 GHz.
  52. cpufreq stats: 3.10 GHz:-nan%, 2.80 GHz:-nan%, 2.30 GHz:-nan%, 1.90 GHz:-nan%, 1.40 GHz:-nan% (101300)
  53. analyzing CPU 4:
  54. driver: acpi-cpufreq
  55. CPUs which run at the same hardware frequency: 4
  56. CPUs which need to have their frequency coordinated by software: 4
  57. maximum transition latency: 4.0 us.
  58. hardware limits: 1.40 GHz - 3.10 GHz
  59. available frequency steps: 3.10 GHz, 2.80 GHz, 2.30 GHz, 1.90 GHz, 1.40 GHz
  60. available cpufreq governors: conservative, ondemand, userspace, powersave, performance
  61. current policy: frequency should be within 1.40 GHz and 3.10 GHz.
  62. The governor "ondemand" may decide which speed to use
  63. within this range.
  64. current CPU frequency is 1.40 GHz.
  65. cpufreq stats: 3.10 GHz:-nan%, 2.80 GHz:-nan%, 2.30 GHz:-nan%, 1.90 GHz:-nan%, 1.40 GHz:-nan% (71912)
  66. analyzing CPU 5:
  67. driver: acpi-cpufreq
  68. CPUs which run at the same hardware frequency: 5
  69. CPUs which need to have their frequency coordinated by software: 5
  70. maximum transition latency: 4.0 us.
  71. hardware limits: 1.40 GHz - 3.10 GHz
  72. available frequency steps: 3.10 GHz, 2.80 GHz, 2.30 GHz, 1.90 GHz, 1.40 GHz
  73. available cpufreq governors: conservative, ondemand, userspace, powersave, performance
  74. current policy: frequency should be within 1.40 GHz and 3.10 GHz.
  75. The governor "ondemand" may decide which speed to use
  76. within this range.
  77. current CPU frequency is 1.40 GHz.
  78. cpufreq stats: 3.10 GHz:-nan%, 2.80 GHz:-nan%, 2.30 GHz:-nan%, 1.90 GHz:-nan%, 1.40 GHz:-nan% (99603)
  79. analyzing CPU 6:
  80. driver: acpi-cpufreq
  81. CPUs which run at the same hardware frequency: 6
  82. CPUs which need to have their frequency coordinated by software: 6
  83. maximum transition latency: 4.0 us.
  84. hardware limits: 1.40 GHz - 3.10 GHz
  85. available frequency steps: 3.10 GHz, 2.80 GHz, 2.30 GHz, 1.90 GHz, 1.40 GHz
  86. available cpufreq governors: conservative, ondemand, userspace, powersave, performance
  87. current policy: frequency should be within 1.40 GHz and 3.10 GHz.
  88. The governor "ondemand" may decide which speed to use
  89. within this range.
  90. current CPU frequency is 3.10 GHz.
  91. cpufreq stats: 3.10 GHz:-nan%, 2.80 GHz:-nan%, 2.30 GHz:-nan%, 1.90 GHz:-nan%, 1.40 GHz:-nan% (72444)
  92. analyzing CPU 7:
  93. driver: acpi-cpufreq
  94. CPUs which run at the same hardware frequency: 7
  95. CPUs which need to have their frequency coordinated by software: 7
  96. maximum transition latency: 4.0 us.
  97. hardware limits: 1.40 GHz - 3.10 GHz
  98. available frequency steps: 3.10 GHz, 2.80 GHz, 2.30 GHz, 1.90 GHz, 1.40 GHz
  99. available cpufreq governors: conservative, ondemand, userspace, powersave, performance
  100. current policy: frequency should be within 1.40 GHz and 3.10 GHz.
  101. The governor "ondemand" may decide which speed to use
  102. within this range.
  103. current CPU frequency is 1.40 GHz.
  104. cpufreq stats: 3.10 GHz:-nan%, 2.80 GHz:-nan%, 2.30 GHz:-nan%, 1.90 GHz:-nan%, 1.40 GHz:-nan% (76089)

m
0
l
a c 473 à CPUs
a c 117 À AMD
November 22, 2013 3:45:42 AM

Weird. So you're still using a version of linux? If you were on windows I would recommend those core parking patches I linked above but w/e.

Have you tried changing the power saving settings around in BIOS? C6, C1E, AMD Cool'n'Quiet etc.? I've been having some issues when overclocking recently so they came to mind, it may make a difference for you. Spikes could be caused by the power saving features working incorrectly. You could also try updating the BIOS to the latest version.
m
0
l
November 22, 2013 7:09:51 PM

Alright, last bit of information in hope that it might help.
Produced an HTML page of my detailed specs, and I saw this: http://postimg.org/image/ej4mfa1jr/
It's the only section highlighted in red. I can't really tell what the issue is, but would a problem with the Serial BUS be something that could cause CPU spikes? Perhaps a missing/defective driver? Altough again, I don't see why it would have uninstalled itself out of nowhere, but who knows.
m
0
l
!