CPU, SERIOUS: Incorrect 128bit (SIMD-SSE) floating point addition

Collapse
X
Collapse
 
  • Page of 1
  • Time
  • Show
Clear All
new posts
Previous template Next
  • Dodiesis94
    Junior Member
    • Jul 2020
    • 4
    #1

    CPU, SERIOUS: Incorrect 128bit (SIMD-SSE) floating point addition

    Hi,

    I recently had various issues on my pc, so I decided to use BurnInTest to check if one component has to be replaced.

    Code:
    SERIOUS ERROR SUMMARY FOR THE LAST TEST RUN
*******************************************
SERIOUS: 2020-07-26 19:22:57, CPU, SERIOUS: Incorrect 128bit (SIMD-SSE) floating point addition (x 1102)
SERIOUS: 2020-07-26 19:23:01, CPU, SERIOUS: SSE operation verification failed (x 90)
    The problem is that I don't know which component should I replace, if the CPU or the motherboard ... I'd rather avoid buying another CPU just for testing purpose.

    I ran the tests with log trace level 1. If they can be useful, I would gladly share the logs but I just can't figure out a way to do that

    Thanks
    Simone
    Tags: None
    • David (PassMark)
      Administrator
      • Jan 2003
      • 11023
      #2
      So the CPU did some floating point maths and got the wrong answer.

      Something like.
      1.5 + 1.5 = 4.1.

      So this can be due to bad RAM or a bad CPU.
      I'd try using MemTest86 next to see if the RAM is bad.

        Comment

        • Dodiesis94
          Junior Member
          • Jul 2020
          • 4
          #3
          Thank you kindly for your fast reply.
          I ran MemTest86 and the results were good, even though it faced some visual oddities (only during some "hammer tests").
          I made some photoes of those oddities, maybe they are typical for this kind of test, can't tell ...
          So could it be the CPU ? Is there a chance that it's the motherboard instead?

          Simone
          Test results There were small white regions appearing and disappearing Here the screen went completely blank, like an old tv
          3 Photos

            Comment

            • Dodiesis94
              Junior Member
              • Jul 2020
              • 4
              #4
              I don't understand ... today to try something different I replaced my graphic card with an older one, ran again BurnInTest on CPU and no error showed up
              Any guess?

                Comment

                • David (PassMark)
                  Administrator
                  • Jan 2003
                  • 11023
                  #5
                  No, strange visual artefacts are not normal during MemTest86.

                  Having a bad video card cause visual artefacts is fairly common.
                  But having a bad video card cause CPU errors isn't common at all.
                  Is the power supply adequate for the machine?

                    Comment

                    • Dodiesis94
                      Junior Member
                      • Jul 2020
                      • 4
                      #6
                      I thought it was, maybe it's not ... It's the only reasonable explanation in fact.
                      My mobo has 2 8-pins power connectors, but since the PSU has got only one of them and theorically the second one is not mandatory, I thought it would have worked. It did for two weeks, then some issues showed up. With hindsight, for a 12-cores CPU the second power connector should be mandatory.
                      I'll have to replace the PSU I guess, It could have been worse.

                      Thanks again, you have been really helpful, and so are these PassMark products.

                      Simone

                        Comment

                        Previous template Next
                        Working...