Page 3 of 3 FirstFirst 1 2 3
Results 21 to 22 of 22
  1. #21
    ROG Enthusiast Array morph. PC Specs
    morph. PC Specs
    Laptop (Model)MacBook Pro 13" i5, Surface Pro X
    MotherboardROG Maximus XII Formula
    ProcessorIntel® Core™ i9-10850K @5.1Ghz
    Memory (part number)G.SKILL 32GB (16X2) 3466MHZ 16-16-16-32-1N TZ RGB
    Graphics Card #1ROG Strix RTX3090oc
    MonitorROG Swift PG279Q
    Storage #1Samsung 970 EVO Plus 1TB NVMe
    CPU CoolerEK Quantum Velocity Nickel Plexi 2x PE360 Radiators
    CaseLian LI PC-O11DXL-W
    Power SupplyCorsair AX1000
    Keyboard SteelSeries Apex Pro
    Mouse Final Mouse Air58
    Headset SteelSeries Arctis Pro Wireless
    Mouse Pad ROG Sheath
    OS Windows 10 Pro
    Network RouterBillion BiPAC 8800AXL
    morph.'s Avatar
    Join Date
    Aug 2017
    Reputation
    10
    Posts
    68

    Great stuff your guide helped coming from an overclocked 8700k to a 10850k Thankyou. What was your SP as a reference?

    I would have loved to see some more settings that should be either enabled or disabled in the bios just to confirm what I did was right.

    More in-depth stuff about your 5.1/5.2 passes would also be greatly appreciated.

    Another observation I've had was when running LLC4,5,6,7 the vdroop was definitely lower at LLC7 than 4 which allowed me to run a lower vcore overall. With such a large gap why was it recommended to use LLC4? Are we really that worried about overshoot?

    With a SP79 10850k @5.1ghz LLC7 on 1.31v in bios translating to 1.314v in windows can't recall my vmin exactly but its approx 1.28v, will reconfirm after I run a few more stability tests. Honestly, I haven't found an unstable point yet so I'm trying to dial down the vcore a bit more to yield lower temps and power consumption during desk/non-gaming workloads.

    Been noticing lately that my 5.1 oc keeps dropping on desktop idle to my avx offset and for the life of me I cannot work out what is causing these avx instructions. Ive tried to close down all unnecessary apps and what not but still no luck so I've kind of resulted in not having an avx offset.

    When I attempted prime 95 small FFT with AVX it hit past 95 degs (with -1 or 2 offset) and triggered thermal throttling. So I stopped the test and this is even though I have a full custom loop running dual EKWB PE 360 rad blocks.
    Last edited by morph.; 10-02-2020 at 01:16 AM.

  2. #22
    ROG Guru: Yellow Belt Array
    Join Date
    Aug 2014
    Reputation
    70
    Posts
    178

    Quote Originally Posted by morph. View Post
    Great stuff your guide helped coming from an overclocked 8700k to a 10850k Thankyou. What was your SP as a reference?

    I would have loved to see some more settings that should be either enabled or disabled in the bios just to confirm what I did was right.

    More in-depth stuff about your 5.1/5.2 passes would also be greatly appreciated.

    Another observation I've had was when running LLC4,5,6,7 the vdroop was definitely lower at LLC7 than 4 which allowed me to run a lower vcore overall. With such a large gap why was it recommended to use LLC4? Are we really that worried about overshoot?

    With a SP79 10850k @5.1ghz LLC7 on 1.31v in bios translating to 1.314v in windows can't recall my vmin exactly but its approx 1.28v, will reconfirm after I run a few more stability tests. Honestly, I haven't found an unstable point yet so I'm trying to dial down the vcore a bit more to yield lower temps and power consumption during desk/non-gaming workloads.

    Been noticing lately that my 5.1 oc keeps dropping on desktop idle to my avx offset and for the life of me I cannot work out what is causing these avx instructions. Ive tried to close down all unnecessary apps and what not but still no luck so I've kind of resulted in not having an avx offset.

    When I attempted prime 95 small FFT with AVX it hit past 95 degs (with -1 or 2 offset) and triggered thermal throttling. So I stopped the test and this is even though I have a full custom loop running dual EKWB PE 360 rad blocks.
    The guide should only be used as a beginning estimate, and the VF points prediction is not perfect. In fact, two chips with the same SP score can overclock far differently, there are SP 63's that clock as well as SP 90's, and SP 8x's that are worse than some SP63's!

    Two chips with the same SP score can have different voltage grace margins, also known as guardbands, and there is no way for any Bios to test for something like this by any sort of VID or programming. The only way to test for this is to actually run a stress test, and I think only one board in existence allows stress testing in the Bios, and that's a Z490 dark (IIRC).

    Another huge thing that can influence how well a chip will overclock is the scaling. Particularly, the V/F scaling between 5.1 and 5.2 ghz. The smaller this is, the better. If your chip has a low SP score but the scaling between 5.1 and 5.2 ghz is still very strong (e.g. 50mv or lower difference on V/F points--the smaller the VF difference, the better the scaling), then that chip may respond to voltage or overclock much better than a chip that has terrible scaling (like 150mv difference between 5.1 to 5.2 ghz). Such a high V/F delta usually means that that chip is "past" its voltage wall. I don't have multiple chips to test nor do I have access to this information; there are those much more fortunate who have many chips to bin, and I am not one of them.

    My ES chip is SP94, but seems to respond like an SP 8x chip (almost identically to I think criskoe's SP84 chip on overclock.net). VF is 1.305v at 5.1 ghz and 1.393v at 5.2 ghz, which is an average delta. I've seen some of the best chips have a 30-40mv delta and the absolute worst I think was either 170mv or 180mv (something like 1.329v to 1.499v).

Page 3 of 3 FirstFirst 1 2 3

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •