The Notebook Review forums were hosted by TechTarget, who shut down them down on January 31, 2022. This static read-only archive was pulled by NBR forum users between January 20 and January 31, 2022, in an effort to make sure that the valuable technical information that had been posted on the forums is preserved. For current discussions, many NBR forum users moved over to NotebookTalk.net after the shutdown.
Problems? See this thread at archive.org.

    GPU Performance Scaling (Testing)

    Discussion in 'Sager and Clevo' started by codral, Jan 20, 2019.

  1. codral

    codral Notebook Guru

    Reputations:
    2
    Messages:
    56
    Likes Received:
    13
    Trophy Points:
    16
    I've just been testing for the last 3 hours or so and have found some interesting results with GPU Performance scaling:

    My PC specs are below, I've got some little chocks I've made up to elevate the back of my PC, (if anyone wants to 3d print them, feel free to PM me or I can post them on Thingiverse, made them based on the thickness and a basic height designed to elevate but minimise slippage, also to allow the CPU to vent properly) and they make a big difference on temps so the CPU and GPU both don't thermal throttle.

    I don't get any performance difference on Automatic or Max Fan (obviously MAX Q is crap).

    On Tomb Raid benchmark the GPU and CPU both are more stable, the GPU sits right on .831 mV for most of the benchmark, only dropping when the load drops off, this with GPU Performance Scaling DISABLED.
    When it's ENABLED I get a lot more variation in the voltage and max clocks for GPU.

    I also get a performance bump in 3dmark Firestrike, from 15100 (enabled) to 15200 (disabled), consistently (with about a 153mhz bump to core and 499 mhz bump to memory).

    With this being said I get no CPU throttling at all either way (-130 mV).

    My theory is with the 180W power supply there's enough headroom that with it ENABLED it seems to scale unnecessarily. Anyone have similar results?

    I have read so much conflicting information on this so figured I'd test it myself.
     
    codral, Jan 20, 2019
    #1
  2. Meaker@Sager

    Meaker@Sager Company Representative

    Reputations:
    9,436
    Messages:
    58,194
    Likes Received:
    17,909
    Trophy Points:
    931
    The max-q power limits will be dictating all this, even a larger brick would not change the behaviour.
     
    Meaker@Sager, Jan 21, 2019
    #2
    joluke and Dr. AMK like this.
  3. codral

    codral Notebook Guru

    Reputations:
    2
    Messages:
    56
    Likes Received:
    13
    Trophy Points:
    16
    Yeah i figured as much myself, and am across the different wattage of models of 1070MQ just thought it interesting that it works better Disabled... And I've read that some of the newer Clevo models of 1070MQ had less max voltage supply, so thought it might help someone get a little boost as I got
     
    codral, Jan 21, 2019
    #3
    Prostar Computer likes this.
  4. djsubtronic

    djsubtronic Notebook Evangelist

    Reputations:
    109
    Messages:
    387
    Likes Received:
    76
    Trophy Points:
    41
    15200 overall score or graphics score? I would think a 100 point difference is within margin of error though, my result fluctuates more than that just running it repeatedly without even changing anything.
     
    djsubtronic, Jan 21, 2019
    #4
  5. Meaker@Sager

    Meaker@Sager Company Representative

    Reputations:
    9,436
    Messages:
    58,194
    Likes Received:
    17,909
    Trophy Points:
    931
    That's true, 100 points is error margin.
     
    Meaker@Sager, Jan 21, 2019
    #5