Intermittent freeze/lock on HTPC, can't pinpoint source

avulhop

Honorable
Nov 16, 2013
5
0
10,510
Here's an image of what I see when it happens. Apologies for quality, as it's from a camera-phone:
iSzMaRHl.jpg


It's very intermittent. Sometimes it occurs a couple times in a day, other times I'll go over week or two without incident, all with no discernible pattern or change is usage behavior.

When the machine locks, it generally starts to stutter badly if it's being used to watch video or play a game. The stutter includes the audio, but once the PC totally locks, the audio cuts out. It's totally unresponsive and requires a hard reset. It's happened when I'm not looking at the screen (idling with screensaver, some hobby cryptocoin mining, etc), so I can't say 100% that there is always a stutter. I've not had this running hot for extended periods of time doing cryptocoin mining. At most, it runs a miner overnight or for a few hours when nobody is using it (this machine is the living room HTPC for a family of 4).

Nothing appears in the Event Viewer other than the typical error messages stating that the previous shutdown was unexpected. Never any smoking gun prior to the lock.

Machine details:

  • OS: Windows 7
    CPU: Intel Pentium G645 2.9GHz
    Mobo: H77MU3 LGA 1155 mATX
    PSU: CX Series CX600M 600 Watt ATX Modular
    RAM: Ballistix Sport 4GB DDR3-1600
    GPU: SAPPHIRE 100355OCL Radeon HD 7850 2GB 256-bit GDDR5 (OC version)
    OS SSD: SSDNow V200 SV200S37A/64G 64GB SATA
    Media HDD: Barracuda 2TB 7,200 RPM SATA
    some SATA DVD burner
    some cheap media card reader
    some cheap USB3.0 hub that slots into PCI
    using a Samsung HD TV as a monitor, connected via HDMI.
    All hardware is factory-default, no overclocking or other mods done.
    Catalyst Driver: 13.11
    AMD Audio Driver: 7.12.0.7717

RAM has been through Memtest86 and passed.
I've used Furmark to do a burn in on the gpu and I've not seen it crest over 60C. Last time I did it (today, 2/8/2014) I never saw a temp above 57C. (I've also never seen a temperature above 52C while running a cryptominer, generally in the 40s.).
CPU never seems to run too warm. I keep a widget on my desktop with the core temps for both cores, and they spend most of their time sub-40C. I've yet to torture them with Prime95, but I've not seen any reason to think it's CPU related yet.
I've turned on Windows Driver Verifier and gave it some light and heavy loads (streaming Hulu, turning on a crypto miner, Furmark while driver verifier was enabled, etc) and nothing triggered a BSOD. When I did get one of the freezes, there was no BSOD or minidump to analyze.
This issue has persisted through a PSU replacement, so it's highly unlikely that power is the culprit.
I've run a HDD tester from Seagate on my "media" HDD and found no problems.
I've run an SSD benchmarker (Crystal Disk Mark) and nothing failed and the numbers seemed reasonable.
All of the important and most of the optional Windows Updates are installed.

I'm having trouble figuring out the source, especially since I can't reliably stimulate the system to reproduce the issue. Any ideas?
 

dmmbbs

Distinguished
Jan 19, 2011
237
0
18,710
Disable everything not needed in the BIOS, remove all peripherals not needed for a minimal boot, do a minimal boot and see if it's a hardware problem. If not, try safe mode/system restore/AMD clean uninstall to see if OS/software/driver corruption.
 

avulhop

Honorable
Nov 16, 2013
5
0
10,510




As much as I'd be willing to do these things, since I can't force the issue to reproduce, I don't think it will be a valuable test. I've had runs of weeks on end where it didn't occur. I'd like to avoid running my family HTPC in a hobbled state for a month before I determine it was or was not the GPU or some other component.

Any ideas on how to pinpoint the issue that's so intermittent? Maybe ideas on how to trigger the problem (i.e., "if it's X then it should happen every time you do Y" sort of things)
 

avulhop

Honorable
Nov 16, 2013
5
0
10,510
Update: While driver verifier was running, I was skyping with family out of town. Shortly after starting the video call, it got extremely stuttery then locked up with similar visual artifacts. The audio did, in fact, hang on a low buzz until I did a hard reset. (again, no BSOD or minidump)

However, during the hard reset, it took forever to post. The mobo splash (American Megatrends, mobo serial number) came up, but it was probably 2 minutes before the "Press <DEL> to enter setup" and single POST beep occurred.

Wondering, could this be a symptom of some damage from the hard reset(s), or a possible culprit showing itself?