Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie
Hi there,
There is an issue with role permissions that is being worked on at the moment.
If you are having trouble with access or permissions on regional forums please post here to get access: https://www.boards.ie/discussion/2058365403/you-do-not-have-permission-for-that#latest

Would there be a difference in OC ability?

  • 22-03-2016 8:14pm
    #1
    Registered Users, Registered Users 2 Posts: 403 ✭✭


    Hi all,
    Was poking around the 380X line, as my mate was upgrading his GPU for a few games this year. I noticed that the majority of the 380X models come with 2 6 Pin Power requirements, but HIS and Gigabyte only require 1 8 Pin connector.

    After a failed attempt at finding any benchmarks or explanation, would the 8 pin hamper the OC ability of the Gigabyte/HIS cards (which work out €10 cheaper for someone on a tight budget) or is it that 2 6 pin = 1 8 pin?
    Many thanks!
    Tagged:


Comments

  • Closed Accounts Posts: 6,934 ✭✭✭MarkAnthony


    AMD Cards don't overclock well under the new drivers as you can't increase core voltage anymore. I'd be surprised if it mattered tbh.


  • Registered Users, Registered Users 2 Posts: 403 ✭✭Eoinmc97


    ASUS models can modify core voltages because they do not use reference boards, instead they use their C768 boards. Howevee, their 380X uses 2 6pins, so I think for the sake of €10, he might aswell just get the Gigabyte card.
    Windforce x2 are very quiet from what I've seen (except when they gave the 390 2 fans lol and a 960 got 3 instead)


  • Closed Accounts Posts: 6,934 ✭✭✭MarkAnthony


    Eoinmc97 wrote: »
    ASUS models can modify core voltages because they do not use reference boards, instead they use their C768 boards. Howevee, their 380X uses 2 6pins, so I think for the sake of €10, he might aswell just get the Gigabyte card.
    Windforce x2 are very quiet from what I've seen (except when they gave the 390 2 fans lol and a 960 got 3 instead)

    The issue is at a driver level.

    The voltage is being applied after the core/mem clocks causing a crash at boot up. AMD are in no hurry to fix it despite being told over and over.


  • Registered Users, Registered Users 2 Posts: 403 ✭✭Eoinmc97


    Ok, thanks for that! I'd say the reviewers at HardOCP didn't restart whilst testing the ASUS model.
    Do you know if the 8 pin would theoretically change OC ability, assuming the voltages could be unlocked? Or is 1 8pin the same as 2 6 pins?


  • Closed Accounts Posts: 6,934 ✭✭✭MarkAnthony


    Eoinmc97 wrote: »
    Ok, thanks for that! I'd say the reviewers at HardOCP didn't restart whilst testing the ASUS model.
    Do you know if the 8 pin would theoretically change OC ability, assuming the voltages could be unlocked? Or is 1 8pin the same as 2 6 pins?

    I have to admit I'm not sure. I'm not sure what a 8 pin v 2 x 6 pin delivers. I'd assume the 2 x 6 pin would allow more power.

    I've just got a modest OC on my 290 now which doesn't use voltage. To be fair I do give it the max power limit.

    All in all not very helpful I'm afraid!

    If you do manage to OC with voltage would you mind letting me know how to fix it? :)


  • Advertisement
  • Registered Users, Registered Users 2 Posts: 403 ✭✭Eoinmc97


    Did some poking around that might have solved my question.
    According to the ATX standard for PSUs, (all current and previous revesions have not changed this) a 6pin PCI-E power connector delivers 75W, and a voltage of 0.863, whilst an 8pin connector doubles these values. A 6+2pin lowers the voltage to 1.655 (which is 0.123 lower than a single 8pin).
    Now, these would both deliver the same power, however it now comes down to the ripple in your PSU over the individual connectors. HardOCP do a good rundown of these in their reviews, so from what I can tell, it starts to come down to the steadiness ofyour current flow.
    VRMs and TTDs and other components on the GPU that regulate power delivery and conversion operate best at stabler currents, as the heat caused in rippling (as energy cannot be lost, it is converted to heat) will cause these components to heat up. These components have a safe cap at 113° when using Nichicon capicitators, and lower depending on other brands. The cooler operating temps with lower ripple improve component longevity and power stability.
    Now, when you overvolt your GPU, whilst you may set a theoretical 1.3V limit, some of this is physically lost due to interference and ripple. Also, the continuous delivery is important to guarantee higher clocks, as voltage switching does not need to occur as rapidly when the same voltage at the same amperage is delivered.


    So in the end, it depends purely on this;
    Your power delivery to your GPU
    Your cooling situation
    Stability of your GPU to effectively maintain set voltage values without loss

    All of these, are way beyond the scope of your average person to compute, so we then fall down to the sillicon lottery.


    In terms of overvolting your GPU, it seems reference boards cannot overvolt yet, but custome boards can. ASUS, Gigabyte, Sapphire and XFX all offer different board designs. However, the 200 series only had ASUS, Sapphire and XFX to offer different board designs. Whether you have one, you would have to use GPU-Z to see, and see if it's reference or not.
    EDIT: Windows 10 quick-boot/fast-boot disabled allows the OC settings applied in MSI Afterburner to stick. This works in Crimson 16.1 and beyond. Hope this helps!

    HardOCP overclocking review of the R9 380X from ASUS: http://www.hardocp.com/article/2015/12/14/asus_strix_r9_380x_directcu_ii_oc_overclocking_review/4#.VvV-LPrfWrU


Advertisement