Jump to content

eldrin

Members
  • Posts

    7
  • Joined

  • Last visited

Posts posted by eldrin

  1. Thanks for the update. Will it also report details about what exactly is the hardware flaw?

     

    Today I've swapped that second 4GB SODIMM with a different one that I ordered. However, the same thing happens with it unfortunately.

     

    Previously I was using a "low density" Crucial SODIMM (with 16 chips on the memory module), but I read that some systems work better with high density memory. So the one I installed today was the high density version, also from Crucial, which uses 8 chips on the memory module. I was hopeful this would work since the 4GB SODIMM that came with the computer also uses 8 chips.

     

    I wonder if it could be some problem with the memory controller itself. Maybe it doesn't work properly in dual channel mode. Or maybe the system needs two identically branded SODIMMs installed, although the specs of the current ones (Crucial and SK Hynix) are exact matches.

  2. A further update: it seems to make a difference whether I have one DIMM installed or two.

     

    The system originally came with just one 4GB DIMM. I've done the only upgrade possible on this system: added a second 4GB DIMM to take the total RAM to 8GB.

     

    After upgrading to 8GB, I had run Memtest86+ for 11 hours, and Prime95, and no errors were reported. Also, Windows has not crashed or seemed unstable.

     

    However, I've now done some further testing with Prime95, running the Blend test. Sometimes, after a few hours, it'll encounter this: "FATAL ERROR: Rounding was 0.5, expected less than 0.4". Checking the Windows system log in Event Viewer, at around the same time I see an error "Windows removed bad memory regions from this PC". It's weird that it only happens sometimes, and only after several hours of Prime95 testing have passed.

     

    So perhaps AIDA64's System Stability Test is encountering a similar problem when 2 DIMMs are installed in this system, but it's doing so much more quickly. If so, surely it would be more helpful if there was an error message of some kind, instead of just stopping silently and then not doing anything?

  3. On one of my machines, if I include 'Stress system memory' in the System Stability Test, it runs for a few seconds and then activity just dies out on its own. The test simply stops stressing my CPU and memory, etc. and just goes into a dormant state. See attached screenshot.

     

    Everything else seems fine with the system. I can run other tests like Memtest86+ and Prime95 for hours without issue.

     

    post-27469-0-49714500-1406195347_thumb.png

×
×
  • Create New...