Best BurnInTest settings for long-term stress testing on a custom-built PC

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • alicesphere
    Newbie
    • Nov 2025
    • 1

    #1

    Best BurnInTest settings for long-term stress testing on a custom-built PC

    Hello everyone,

    I’m new to the forum and just started using PassMark BurnInTest to validate the stability of a newly built custom PC before putting it into regular use. I’d really appreciate your input on configuring an effective stress test setup.

    Here’s my current hardware:
    • CPU: AMD Ryzen 7 5800X
    • GPU: NVIDIA RTX 3070
    • RAM: 32 GB DDR4
    • Storage: NVMe SSD + HDD
    • PSU: 750W Gold

    My goal is to run a long-duration stress test (12+ hours) to catch any early-life hardware issues. However, I’m not sure which combination of tests and settings will give me the best coverage without causing false positives (e.g., overheating artifacts vs real faults).

    Some specific questions:
    1. Which modules should I enable? Should I test CPU, memory, disk, GPU all at once or run them sequentially?
    2. Test duration: Is 12 hours adequate or should I aim for 24+ hours?
    3. CPU/GPU load level: What percentage load is safe yet stressful? Should I leave headroom or push to 100%?
    4. Monitoring temperatures: How should I set temperature thresholds to avoid thermal shutdowns while still stressing the hardware?
    5. Interpreting errors: When BurnInTest reports “soft errors” (e.g., memory errors), how likely is it that these indicate real hardware issues versus just random transient glitches?

    Thanks in advance for any advice or pro tips. I want to make sure this PC is stable and reliable before I use it for work. Looking forward to learning from your experiences!
  • David (PassMark)
    Administrator
    • Jan 2003
    • 11085

    #2
    This page provides some guidance.
    https://www.passmark.com/support/bur...les.php​

    12h should be enough unless
    1) The PC is to be used with something critical (medical equipment, space applications, air traffic control systems, etc..)
    2) You hard drives are enormous and/or slow.

    Test them all at once.
    Try and test in the 90% to 100% load range for all components (as indicated in Task manager). It won't be possible to get smooth load, but you don't want too little, but nor do you want to attempt to go past 100%. See link above.

    If your cooling is OK, there should be no shutdowns. Many components now throttle in any case.

    For a 12h test the change of getting "soft errors" is low. So any error should be considered to be real.


    Comment

    Working...