HP OMEN 15 and 17 owners lounge

Discussion in 'HP' started by tweake628, Aug 23, 2017.

  1. OMEN 17-cb0009nl

    OMEN 17-cb0009nl Notebook Enthusiast

    Reputations:
    0
    Messages:
    48
    Likes Received:
    8
    Trophy Points:
    16
    Thanks for the reply but I already put OGH UV feature to off (I use TS for that purpouse).
    Still getting the strange red PL1 on TS limits monitor after 20 second. Before that, CPU is using 69.5W PKG Power and 3900 mhz all 6 cores speed (with good temperatures under 80°C)
    After about 20 seconds PKG drops to 45 W and CPU speed to 33-34 multiplier (tempo go to 60-65°C).

    OGH is not running (not even in background since I killed the app in task manager).
    So its either something I am not doing right in TS setup (maybe an option under FIVR or TLP) or there is a 45 wattage limit for CPU on our Omen cb0 with F.35 bios after 20s of full core speed?

    Also, If I try a game (example Star Citizen), Riva Tuner shows 3900 mhz stable on CPU. What gives? It goes down only on synthetic tests? I don't understand....

    Do you know what are the default good options to check/unckeck under FIVR and TLP so I may test them?
     
  2. Medariuz

    Medariuz Newbie

    Reputations:
    2
    Messages:
    9
    Likes Received:
    2
    Trophy Points:
    6
    as a personal rule, I don't modify any parameter that TS reads from the settings imposed by HP, I only modify the UV under TPL, these are my settings, and I don't have any PL problem under heavy loads (synthetic benchmarks) and combined loads (gaming). Also, I use the windows power slider in performance, and OGH in performance profile, auto fans.

    edit: as an anecdote, I noticed that if I changed the prochot offset to something higher than 97°C, the clocks dropped and a PL1 of 45w would be applied forcefully, regardless of your TPL settings or power/performance profile.
     

    Attached Files:

    OMEN 17-cb0009nl likes this.
  3. OMEN 17-cb0009nl

    OMEN 17-cb0009nl Notebook Enthusiast

    Reputations:
    0
    Messages:
    48
    Likes Received:
    8
    Trophy Points:
    16
    Well that did it @Medariuz. Your screenshots were quite useful thanks.
    I was already on performance OGH, but not in windows performance (thou I always thought the OGH performance would automatically put the windows one into performance mode too). My bad.

    Now the CPU is rock solid on 4000 mhz and the cinebench R32 is much faster (score of 7948), but the peak temperature on CPU is unacceptable on the long run for me with your wattage settings (80-90W) even with the recent MX-4 repaste I was looking at peaks of 90-94°C during test run (no thermal troooteling thou, CPU can take it apparently).
    Setting all 6 cores down from x40 to x39 (3900 Mhz) the score is only marginally worse in Cinebench (7736) but the CPU temps are in the mid 80's (82-86 depends on cores). See attachments.


    Just out of curiosity, do you undervolt your GTX using MSI afterburner too? I just started using it and RTX 2070 I've set to +750 mhz MEM, GPU is standard speed, and voltges I set up as per picture i've attached. Do you have any raccomandations on that one too? GPU runs pretty cold durng test and games (about 55°C)
     

    Attached Files:

  4. Medariuz

    Medariuz Newbie

    Reputations:
    2
    Messages:
    9
    Likes Received:
    2
    Trophy Points:
    6
    nice!

    yep, I run oc scanner about a year ago, and after several runs got a custom curve that I flattened at 1770 mhz. I noticed that if the clocks got higher, many games would crash. Until now I've been able to play and finish: Control, Cyberpunk 2077, Witcher 3 and many others with this curve (see attachment).

    Regarding the GPU memory, after several tests using Superposition benchmark and unigine heaven, I noticed that there's a drop in average frames and final score if I OC the memory clock. and that makes sense, because AFAIK and read around the web, the power budget is quite limited in (at least mine) a RTX 2070 max-q, mine is 90w max boost. so I left the memory clock at +0 or stock


    By the way, I highly recommend repasting with Thermalright TFX, I repasted recently because Cyberpunk was frying this laptop. not anymore :D.
     

    Attached Files:

    OMEN 17-cb0009nl likes this.
  5. OMEN 17-cb0009nl

    OMEN 17-cb0009nl Notebook Enthusiast

    Reputations:
    0
    Messages:
    48
    Likes Received:
    8
    Trophy Points:
    16
    Ah ok, I have the full 2070 (115W I belive it is) so I shoukdn't probably copy-paste your results. I will try going up 50-100 mhz since mine is runnins so stable and cool it probably has some headroom.

    As for thermalright TFX it is on my list for the near future. I will test the MX-4 for a month or so more probably under stress and for the next repaste I'll go with your raccomandation thanks!

    As of curiosity, what temps do you get on your 9750H with that paste during a Cinebench R32 multicore run at x40 cores (4000 mhz?)
     
    Last edited: Feb 8, 2021
    Medariuz likes this.
  6. OMEN 17-cb0009nl

    OMEN 17-cb0009nl Notebook Enthusiast

    Reputations:
    0
    Messages:
    48
    Likes Received:
    8
    Trophy Points:
    16
    addendum on the last post: after setting to 'performance' mode the GPU is automatically OC's and goes up to 1770 mhz (with the limit being the wattage in Rivatuner statistics, not heat). Actualy I'm surprised how cool the GPU die remains even under load. Must be because the GPU die is twice the size of the Intel CPU die, making more contact with the heatsink and dissipating better?
    I wish Intel die was also bigger, maybe it would enable us to use the full x40 an all 6 cores all the time.
     
  7. seanwee

    seanwee Father of laptop shunt modding

    Reputations:
    422
    Messages:
    1,693
    Likes Received:
    816
    Trophy Points:
    131
    Yes, its due to thermal density. The intel cpu is more thermally dense than the gpu which is why its so much harder to cool.

    This is why using liquid metal vs even a high end thermal paste on the cpu still yields a 5-10C advantage.
     
  8. Burrick

    Burrick Notebook Geek

    Reputations:
    10
    Messages:
    91
    Likes Received:
    7
    Trophy Points:
    16
    This is a result of the new OGH, not F35. Checkout the other two threads in the HP forum:
    bios update for hp omen 17t cb-0000 F35
    OMEN command center causing core 0 temp spikes on laptop
    I got it back to normal by disabling OmenCAP and XTUservice. OGH uses those two services to regulate the CPU relative to the performance mode buttons. The machine seems to boot in Performance by default which is good, 90W/90W in my case with the i9-9880h, and then historically it stayed there in performance and the other two dialed it down. The new OGH adds a 28sec timer to Performance which is wretched. Cripple it and use ThrottleStop to manage that stuff, including undervolt since OGH pairs core/cache.
    Now that I'm using Notebook Fan Control which goes to 100% early enough that I don't need manual fan, I only keep OGH around for the keyboard lighting control. Light Studio obviates that even, but it's a CPU hog so waiting for an improved version.
     
  9. OMEN 17-cb0009nl

    OMEN 17-cb0009nl Notebook Enthusiast

    Reputations:
    0
    Messages:
    48
    Likes Received:
    8
    Trophy Points:
    16
    I made a test.
    If I disable both OmenCAP.exe and XTUservice services, my notebook goes into automatic reboot by itself.
    If I disable only OmenCAP.exe, nothing happens apparently.
     
  10. Burrick

    Burrick Notebook Geek

    Reputations:
    10
    Messages:
    91
    Likes Received:
    7
    Trophy Points:
    16
    That's really weird. Do you have XTU itself installed?
     
Loading...

Share This Page