MSI GT73VR Safe & Stable Overclock

Discussion in 'MSI' started by RalfTech, Jan 18, 2018.

  1. Tokyotim

    Tokyotim Notebook Enthusiast

    Reputations:
    24
    Messages:
    20
    Likes Received:
    13
    Trophy Points:
    6

    Sorry about that I've never used any forum before so I didn't realize that was an annoying thing. Mr. Falkentyne I used a Prime blend test.
     
  2. Falkentyne

    Falkentyne Notebook Virtuoso

    Reputations:
    3,916
    Messages:
    3,712
    Likes Received:
    4,787
    Trophy Points:
    331
    Phoenix,
    he said hes sorry and shows remorse. Seems genuine but don't let it happen again.

    Ok here's your blend test at your exact settings and undervolt with liquid metal cooling.
    prime_4.2ghz_adaptive_blend_autollc_undervolt.png

    True but there is actually one VERY good use for prime95, provided you disable the CPU power virus stuff first:
    It can be used as a very reliable RAM, IMC and VCORE test, rather than as a "how good is your heatsink?" test:

    1) disable FMA3 and AVX.
    2) Prime with custom settings:
    min FFT: 512K, max FFT: 4096K, total RAM to use: 75% of total system RAM in MB value
    (for 32GB, this is 75% of 32768, or 24576)

    Different FFT sizes will test different things. 1344K tests CPU Vcore, 2048-4096K tests RAM or CPU IMC (I don't remember which is which, you can google it somewhere its been mentioned).

    This is a good way for testing RAM overclocks, if you can pass UEFI Memtest86 without errors (passing memtest86 first means you won't destroy your Operating system by trying to boot into it, but doesn't mean your RAM tweaks are fully stable).
     
    Phoenix and KY_BULLET like this.
  3. Papusan

    Papusan JOKEBOOKS = That sucks!! STAHP! Dont buy FILTH...

    Reputations:
    11,151
    Messages:
    17,968
    Likes Received:
    26,376
    Trophy Points:
    931
    Maybe try Cinebench R15 and post results. One with HWiNFO64 and one without for highest possible score. A real world workflow:)
    No need for running P95 whatsoever you say with or without FMA3, AVX load. Aida64 tests cover the whole range. You can even test out AVX, AVX2 and FMA workflow for your processor (Not the massive load as in P95, but still more than enough for verifying this type load for your system). And AIDA64 workload is used everywhere to validate hardware components.
     
    Last edited: Feb 7, 2018
    Phoenix and KY_BULLET like this.
  4. Falkentyne

    Falkentyne Notebook Virtuoso

    Reputations:
    3,916
    Messages:
    3,712
    Likes Received:
    4,787
    Trophy Points:
    331
    @Papusan
    can the TRIAL version of AIDA64 do this?
    I thought the trial version had limitations.

    Actually if you're tweaking RAM tertiary timings, you should do the prime test I mentioned, AIDA, Asus Realbench, and other tests mentioned on overclock.net (like HCI memtest and others)
    You can't rely on one test when tweaking subtimings on RAM. One "1" turns into a "0" randomly and your OS can be gone!
     
  5. Papusan

    Papusan JOKEBOOKS = That sucks!! STAHP! Dont buy FILTH...

    Reputations:
    11,151
    Messages:
    17,968
    Likes Received:
    26,376
    Trophy Points:
    931
    As I said in previous post... AIDA64 workload is used to validate hardware components in systems everywhere (both for desktops and notebooks). Of course stress tesing with different workflow is better... But P95 ain't needed!! Even idle load is a form for stress test when you shall validate your overclock/stability.
     
  6. Falkentyne

    Falkentyne Notebook Virtuoso

    Reputations:
    3,916
    Messages:
    3,712
    Likes Received:
    4,787
    Trophy Points:
    331
    Prime95 is fine if used properly.
    Even unclewebb agrees.

    http://www.overclock.net/forum/18051-memory/1630388-comprehensive-memory-overclocking-guide.html

    Can we just agree to disagree sometimes? We are different people. Won't always agree on the same thing.
    This isn't small FFT blow up the CPU testing I'm talking about here....
    AVX and FMA are DISABLED.

     
    hmscott likes this.
  7. Phoenix

    Phoenix Super Tweaker

    Reputations:
    12,465
    Messages:
    17,164
    Likes Received:
    23,486
    Trophy Points:
    931
    no problem. I've unblocked you. Sorry if I sounded rude I am not usually, but you see, when you tag someone with the @ sign, they get a notification. that's all well and good for example when you are posting something that requires the attention of that particular member, but doing it in every single post where there is not really anything that requires the immediate attention of that member and twice in the same post is what gets annoying

    Thanks for understanding.

    Cheers
     
    Vistar Shook and KY_BULLET like this.
  8. Tokyotim

    Tokyotim Notebook Enthusiast

    Reputations:
    24
    Messages:
    20
    Likes Received:
    13
    Trophy Points:
    6
    I replied to the wrong post but here is the Cinebench score you mentioned. Fans on auto. I have read about liquid metal and it seems a little intimidating to a new user like myself. I have only opened up this laptop to replace the SSD

    Cinebench.jpg
     
    Last edited: Feb 7, 2018
  9. Tokyotim

    Tokyotim Notebook Enthusiast

    Reputations:
    24
    Messages:
    20
    Likes Received:
    13
    Trophy Points:
    6
    Roger that I did not know that I apologize I just saw everyone doing it and thought its what the cool kids did.
     
    Phoenix likes this.
  10. Papusan

    Papusan JOKEBOOKS = That sucks!! STAHP! Dont buy FILTH...

    Reputations:
    11,151
    Messages:
    17,968
    Likes Received:
    26,376
    Trophy Points:
    931
    Please, run without Hwinfo and Anti Virus software for to see if you get a more normal score. The score is Crippled for 4.2GHz.
    792cb is barely better than a stock clocked 7820Hk running 3.5GHz on all 4 cores.
    upload_2018-2-7_22-52-7.png
     
    KY_BULLET and Falkentyne like this.
Loading...

Share This Page