A question for a fellas with AMD b350 socket

Discussion in 'Processors and motherboards AMD' started by Deleted member 268800, Nov 3, 2019.

  1. Astyanax

    Astyanax Ancient Guru

    Messages:
    17,035
    Likes Received:
    7,378
    GPU:
    GTX 1080ti
    I'm fully aware of that, its not just the gpu clock domains but also the pci-e clock generator and mobo's that implement lane divising for power saving.

    Here's what im referring to
    https://www.reddit.com/r/nvidia/comments/dluls0/an_update_mpw90_is_right_about_intel_processors/

    and then this guy...

    https://www.reddit.com/r/nvidia/comments/dj6iil/a_comment_on_nvidia_drivers_on_windows_10_with/

    who's issue is probably that defective by design kingston SSD that he has lied about testing without, if he removed that there'd be no issues with the system.....

    blaming the nv drivers when the system is already experiencing disk related dpc spikes...... lol
     
  2. MelodyUT

    MelodyUT Guest

    Messages:
    6
    Likes Received:
    6
    GPU:
    GTX 1050 Ti / 4 GB
    I see, that's some good information.

    However, when devices are using C States / Power Saving, the routines of the drivers hold everything for a bit longer, I think that's generally normal. I think it's normal to get, let's say, 200 μs latency from NVIDIA driver when the GPU is downclocking down to, let's say, 300 MHz clock and 400 MHz memory, plus Active State Power Management (ASPM) for PCI-E also limits link speed and active lanes.

    Of course, when the driver spikes to absurd values then it's not a power saving issue anymore. In order to be 100% sure you need to run all your devices at maximum clock and then test with LatencyMon. If you still get spikes, then something's definitely up.
     
    Deleted member 268800 likes this.
  3. 386SX

    386SX Ancient Guru

    Messages:
    2,084
    Likes Received:
    2,243
    GPU:
    AMD Vega64 RedDevil
    I guess your RAM is fine? No errors in Testmem5 or other tools? I ask because since weeks I fiddle with timings and stuff and observed unbelievable strange stuff happening, starting from stutters over lags to occasional hangs, even with setups which were considered stable.
    What values do CPU-Z, RTC and AIDA cache benchmark show you? Does your system pass 20 runs of Testmem5? Almost all configuration which stuttered or took far too long to load anything threw errors on cycle 10+.

    Just an idea ... I hope it helps.
     
    Deleted member 268800 likes this.
  4. Hello bro, thank you for being care about my situation, and yes in the pass i did run like woah 12-15 hour test on each dimm, no errors, but i'm not having stutter at all just the visual lag, and game not looking smooth and some input lag too :/

    I think my problem is 100% electricity but also this motherboard with the root poorts and the hpet issues aff is killing me, hope i can move from Venezuela, this situation is much worse every day
     
    386SX likes this.

  5. 386SX

    386SX Ancient Guru

    Messages:
    2,084
    Likes Received:
    2,243
    GPU:
    AMD Vega64 RedDevil
    OK, RAM ruled out. :)

    To rule out an electricity problem I would take the computer to another house or company, from which I know they got clean power. Some UPS may help there, right.

    But another thought comes to mind. I had an issue once with a timed breaker which flipped not only the power of the device connected to it but introduced a spike into all devices in the chain (back to the main fuse), resulting in a sudden black screen.
    May it be possible the malicious electricity is introduced from something out of your household? Fridge maybe? Old devices in general?

    Do other computers behave strange, too? Or is this computer the only one?

    Tbh. I never heard of stuff like that. But nevertheless good luck and heads up! ;)
     
    Deleted member 268800 likes this.

Share This Page