Predator Triton Qames run in slow motion (5-10fps) after CPU/GPU repaste? Triton 500 (51-515)

agentofrandom
agentofrandom Member Posts: 20

Tinkerer

edited January 2021 in Predator Laptops
Hello everyone,
I have a Predator Triton 500 (51-515) which I recently repasted per suggestion on this forum to deal with game performance issues. 

This is a tricky laptop to repaste, as you have to completely disconnect, remove, and flip the MB to access the heatsink and CPU/GPU. Nevertheless, I was able to do it, and after reconnecting everything I am seeing an immediate improvement in thermals, especially for the CPU (used to idle at 85-90 degrees!)

But everything changes when I open a game - games now run in slow motion, around 5-10fps. My first guess was that they were running off the integrated Intel GPU, but I checked and the NVIDI GPU is still detected and the default. Thermals while running the game still look great, between 70-80 degrees (almost always pushing 95 before), so I don't think it's a heat issue.

I clearly must have messed something up when re-assmbling the laptop, but I am unsure of what, as all of the components related to performance were untouched, other than the CPU/GPU being cleaned and repasted (using 99% isopropyl alcohol which is widely recommended).

Anyone have suggested on what I could have touched/changed that might cause this? I am pretty new to this, and hoping I didn't damage something!

Best Answer

  • busu3
    busu3 Member Posts: 552 Seasoned Specialist WiFi Icon
    edited January 2021 Answer ✓
    Problems were not there before the repaste, so that must be it...I watched a lot of videos about repasting before doing this, but all the examples included thermal pads that didn't need to be replaced, so the blue paste being something I had to deal with didn't even occur to me. Hopefully I didn't do any permanent damage.

    I am going to order some thermal pads and do the whole operation over again, this time replacing the blue paste with pads. Thanks for your help! I'll probably report back if that solves it.
    Oh yes you have to put the thermal pads back or atleast some sort of paste on them. Good luck on the repaste!

Answers

  • agentofrandom
    agentofrandom Member Posts: 20

    Tinkerer

    Just to update, I tried uninstalling/reinstalling drivers, checked power mode settings, and also tested without my external monitor disconnected just to weed a few things out.

    Task Manager shows CPU/RAM in normal loads, with GPU Utilization at 100% (that seemed odd, but after googling it, it is apparently normal?)

    I used Assasin's Creed Oddysey to do some testing, and the FPS graph visually shows quite an intense shift happening. It's almost like something is switching on and off:



    Still hoping someone has some clues to what might be happening. Thanks!
  • busu3
    busu3 Member Posts: 552 Seasoned Specialist WiFi Icon
    Can you run something like heaven benchmark and see the fps there? Monitor the clock speeds using something like msi ab. Also the laptop opens from the front?(like screws are on the keyboard?) I dont understand why u had to remove the whole MB?
  • agentofrandom
    agentofrandom Member Posts: 20

    Tinkerer

    Thanks for the reply! RE opening the laptop, the screws are on the bottom, so you have to remove the bottom, fully disconnect the MB, and then flip it to access the CPU/GPU/Memory/Storage. Might be easier to look at this video, which is what I followed when dissasembling:

    https://www.youtube.com/watch?v=6-PmEaytJOA

    I downloaded both Afterburner and Heaven Benchmark, and I am definitely seeing the same thing with FPS as in AC:O, where the FPS runs normal/high for a few seconds, then dips low. Usually going between 50-60FPS and 10-15FPS. 

    I am new to this stuff, so I am not 100% sure what I am seeing, but here's what MSI AB is showing while this is happening:





    Hoping this will give some clues as to what is going wrong! Thanks!
  • busu3
    busu3 Member Posts: 552 Seasoned Specialist WiFi Icon
    Its a bit difficult to see the actual values in these graphs since there are no markings. U run heaven benchmark and enable the on screen display of msi ab. There just monitor the GPU clock speeds. From this graph I noticed that your gpu is not clocking beyond 300 or so(i may have seen some other graph so just check).
  • agentofrandom
    agentofrandom Member Posts: 20

    Tinkerer

    Thanks again. You can see the min/max values on the left, so it does look like the core clock is shifting between 300 and 1400 mhz. I’ll check the AB main window to confirm.

    also, I noticed the GPU temp limit is going between on/off a lot, despite the max temp only being around 75 degrees. Is it possible the temperature gauge is wrong, and this is just the GPU is overheating and throttling? It’s totally possible I messed up repasting the GPU, it’s just weird that the temperatures appear to be fine in MSI/PredatorSense
  • busu3
    busu3 Member Posts: 552 Seasoned Specialist WiFi Icon
    You can verify the temps by installing hwinfo. If gpu is throttling then temps would be reaching 95+ for it to drop by 1000mhz. I dont see why it is shifting between 300 and 1400 mhz. U can download throttlsestop and check for any throttling limit since that wont be visible in hwinfo.
  • agentofrandom
    agentofrandom Member Posts: 20

    Tinkerer

    Okay, so the HWINFO shows the same temps as Afterburner, and it also shows the "GPU Temp Limit" going on several times, but I don't know what that really means (might be a red herring).

    I downloaded Throttle Stop, and the Limit Reasons are showing "EDP Other" and "PT2" going red under the CPU section, and EDP Other also Red under the Ring section. There is also the word "POWER" flashing red next to the LIMITS button briefly. Took a video here (sorry it is wobbling):

    https://www.dropbox.com/home/Personal?preview=IMG_16981083.MOV

    I don't know what "PT2" or "EDP Other" means, but now I am wondering if this is a CPU issue, since nothing seems to appear under the GPU heading during the errors.

    Hopefully this helps pinpoint the problem. Thanks again!
  • busu3
    busu3 Member Posts: 552 Seasoned Specialist WiFi Icon
    The GPU temp limit refers to the limit 1 reached by the gpu making it reduce clocks. Just check how much the clocks are before and after this limit is reached. It should be like 1400mhz-->1380 or so(small drops in frequency not something like 1000 mhz at once). PL2 is power limit throttling during turbo and edp other means current limit being reached. Disable turbo boost of your cpu and check again. Can you also run user benchmark and post the results here?
  • agentofrandom
    agentofrandom Member Posts: 20

    Tinkerer

    Thanks! Okay, so I tried turning off Turbo Boost, and that cleared the Throttle Limits going Red, but the same issue with the GPU going back to 300mhz is happening, and same effects in-game, so I think the throttle limits weren't causing the main issue.

    I ran User benchmark and it showed everything normal except the GPU running at 24%:

    UserBenchmarks: 
    CPU: Intel Core i7-8750H - 84.1%
    GPU: Nvidia RTX 2080 (Mobile Max-Q) - 24.4%
    GPU: Intel UHD Graphics 630 - 5.6%
    SSD: Intel Raid 0 Volume 1TB - 166.3%
    RAM: Hynix HMA82GS6CJR8N-VK 2x16GB - 85%
    MBD: Acer Predator PT515-51


    So, question is, what could I have done to the GPU to cause this while re-pasting? Is it maybe a bad pasting job, or could I have disconnected something and not properly reconnected it? 
  • agentofrandom
    agentofrandom Member Posts: 20

    Tinkerer

    Not sure if it’s relevant, but One more thing I’ll mention is that the heat sink had a blue goo/paste everywhere, and not many thermal pads like I was expecting. I took a picture when I was doing the repaste:


  • busu3
    busu3 Member Posts: 552 Seasoned Specialist WiFi Icon
    Not sure if it’s relevant, but One more thing I’ll mention is that the heat sink had a blue goo/paste everywhere, and not many thermal pads like I was expecting. I took a picture when I was doing the repaste:


    It could be related to the repaste u did. Was the problem there before you did the repaste? The blue stuff is paste I hope you replaced them with thermal pads. Also did you make sure the thermal pads are making contact with the heatsink? If your memory modules are heating the same problem can happen. Check Hwinfo for gpu memory temperature(if you are lucky there may be a memory module sensor).
  • agentofrandom
    agentofrandom Member Posts: 20

    Tinkerer

    Problems were not there before the repaste, so that must be it...I watched a lot of videos about repasting before doing this, but all the examples included thermal pads that didn't need to be replaced, so the blue paste being something I had to deal with didn't even occur to me. Hopefully I didn't do any permanent damage.

    I am going to order some thermal pads and do the whole operation over again, this time replacing the blue paste with pads. Thanks for your help! I'll probably report back if that solves it.
  • busu3
    busu3 Member Posts: 552 Seasoned Specialist WiFi Icon
    edited January 2021 Answer ✓
    Problems were not there before the repaste, so that must be it...I watched a lot of videos about repasting before doing this, but all the examples included thermal pads that didn't need to be replaced, so the blue paste being something I had to deal with didn't even occur to me. Hopefully I didn't do any permanent damage.

    I am going to order some thermal pads and do the whole operation over again, this time replacing the blue paste with pads. Thanks for your help! I'll probably report back if that solves it.
    Oh yes you have to put the thermal pads back or atleast some sort of paste on them. Good luck on the repaste!
  • agentofrandom
    agentofrandom Member Posts: 20

    Tinkerer

    Thanks for the help, confirming that replacing the blue paste on the other chips with thermal pads solved this issue!
  • busu3
    busu3 Member Posts: 552 Seasoned Specialist WiFi Icon
  • This content has been removed.