RX Vega Owners Thread, Tests, Mods, BIOS & Tweaks !

Discussion in 'Videocards - AMD Radeon' started by OnnA, Aug 15, 2017.

Thread Status:
Not open for further replies.
  1. mtrai

    mtrai Maha Guru

    Messages:
    1,183
    Likes Received:
    374
    GPU:
    PowerColor RD Vega
    We are actually using what HWBOT submission rules allow.
    http://hwbot.org/news/9039_application_52_rules/
    https://hwbotnews.s3.amazonaws.com/wp-content/windows8-81-10.png
    Allowed optimisations
    • LOD
    • Change tessellation level in graphics driver
    Disallowed tweaks/cheats
    • Altering benchmark files or the rendering
    • Any software or human interaction altering the perceived speed of the benchmark program, tricking it to believe it ran faster
    • Lucid Virtu MVP
    • Mipmap
    • Change benchmark settings
    Tessellation can be adjusted at the driver level on both AMD and Nvidia. So there is no TRICKERY. We are showing stuff. Actually I did not use any trickery I know ie no tricks, no setting priority, no closing services, etc...and a my win 10 install has gotten a bit dirty as well this was the only driver tweak that HWBOT allows. The other things I mentioned are also allowed but did not do any of them. As most know I do things based on real world. I am using a PowerColor Red Devil Vega 64 Air with the PC LC bios flashed, but I use a soft power play table I have created.

    I am still working with my overclock with the new drivers. I am not so sure of needing an 850+ PSU as my gpu max power draw with this load is only 293...but I did adjust some other settings that overdriventool cannot with this other tool. This is with the new driver and using the other tool which changes amperes draw, but I do have a 1050 evga gs PSU.

    Most of the time my drivers will not be approved...hence the screen shot...when I am using public drivers I always post 3dmark links to show.

    If people have an issue with that...so be it. I have been very busy for a while with testing stuff to be able to post new stuff.

    Also I am running my ram at only 3200 right now...tracking down an issue with last C7H wifi beta bios. I needed to elemenate it as the issue.
     
    Last edited: Dec 15, 2018
  2. NAZ2222

    NAZ2222 Master Guru

    Messages:
    420
    Likes Received:
    27
    GPU:
    XFX MERC 6800XT 16G
    OnnA would I gain something by updating my 64 NITRO+ wich has bios ending in .0000, to the one you have in the OP .0774 ? thanks ;)
     
  3. slicer

    slicer Member Guru

    Messages:
    140
    Likes Received:
    51
    GPU:
    Sapphire Vega 64
    Can You make a quide, how to exactly find the suitable clocks and pin it to first page of this thread?
    Or maybe somebody else who has done it before? What games to test and how exactly find the right voltages/mhz for each states...
     
    Jackalito likes this.
  4. slimmy427

    slimmy427 Active Member

    Messages:
    93
    Likes Received:
    19
    GPU:
    EVGA Hybrid 1080ti
    Thanks for all the replies! I have RMAed one card already...it did not go well! The first card suffered from the same exact issues as the replacement card. I can manage to keep the temps low enough with undervolting but, I have that glitch where is gets stuck at 1.05 volts. I also have a 80mm 4000 rpm fan with rubber bumpers sitting on top of the GPU but, its as if Gigabytes back plate with all the heat pipes and thermal pad do nothing to cool the card.
    [​IMG]



    Heres the condensed drama with Gigabyte. Sent them my RMAed card that was overheating and generating a very unnatural voltage chatter...it was so bad you could here average coil whine just moving the mouse courser. I payed for cross ship for a replacement, in hopes that I would not get a lemon or my lemon back. The replacement came and the thermal pads were not even seated on the shunts...there was a thermal pad literally laying across the copper heat pipes. After contacting service for the replacment card, they strung me along with send a replacement for it. They then refused my RMA claim for the original saying it was modified. I asked for proof and a copy of their warranty and they did not provide either. They then charged my card $500 dollars. Told them there is no way I can trust a company to send my card back in working condition or to except the replacement with their poor QAQC and not charge me for that.

    Basic out come, I have a small claims against Gigabyte since they are in violation of the magnison moss act in three different ways and a dispute on my credit card. So I physically can't do anything to these cards until the claims are resolved. I would personally love a refund for all of it but, I imagine the District Justice will rule that they replace the replacement and pay for the filing fees and credit card charge.
     

  5. JonasBeckman

    JonasBeckman Ancient Guru

    Messages:
    17,564
    Likes Received:
    2,961
    GPU:
    XFX 7900XTX M'310
    They refused? Well hopefully that's going to end quickly but that's not what I would have expected especially these days when you can just photo or film the process from acquiring the package to opening it and checking the status of the product which would pretty much be irrefutable for the problem being on their side.

    I know Gigabyte's Vega cards have had some criticism but that's far beyond just the hardware where support is trying to outright shift blame to you and even try charging for it.

    EDIT: Some retailers can be a bit reluctant but that the manufacturer itself would be that bad is pretty baffling risking publicity and losing customers or how to say it, pretty frigging stupid way for them to try and act like that really.
     
    Last edited: Dec 16, 2018
    Jackalito and slimmy427 like this.
  6. slimmy427

    slimmy427 Active Member

    Messages:
    93
    Likes Received:
    19
    GPU:
    EVGA Hybrid 1080ti
    I was really hoping that Gigabyte would see it that way also. It made me wonder if their RMA department is even part of Gigabyte...like its being operated and managed by some company that gets a percentage if they don't fix it (the vega cards). After this started, I went on a google search and there were so many people that just rolled over and took this same response from gigabyte for their RMAs. Seems its been going on for awhile but, for some reason the visibility is being over shadowed by something else.

    I've had trouble with their motherboards also...bios issues...as I assume the parts and bios are to blame with their vega series. Whoever they are contracting to code the bios is not worth it.
     
  7. G13Homi

    G13Homi Member Guru

    Messages:
    112
    Likes Received:
    37
    GPU:
    RTX 4090
    im sorry to here that, but as a new customer to Gigabyte, also my last time ( Aorus Mobo) i can say they have maybe the worst customer support i have ever had the pleasure of dealing with. My issue was no where near the dread your dealing with though.
     
  8. slimmy427

    slimmy427 Active Member

    Messages:
    93
    Likes Received:
    19
    GPU:
    EVGA Hybrid 1080ti
    Thanks, if I had the time I would try and spread the stories to tech publishers but, I just want it over with. I think most companies have that in mind when they start screwing people...capitalism at its worst.

    Onna, what is your clock during the bench? The only way I see clock increase above 1450mhz is to increase the power %. My best graphics score on fire strike is 25000...thats at 50%+ 1590 / 1070. Anything higher than that the system crashes
     
  9. OnnA

    OnnA Ancient Guru

    Messages:
    17,851
    Likes Received:
    6,741
    GPU:
    TiTan RTX Ampere UV
    My GPU is Vega XTX LC AMD reference.
    Yours is AIB Sapphire (IMO You don't want to touch this)
    Nitro is Great GPU BTW (The one of the Best Vega is made by Sapphire and XFX and PowerColor ;) It's ATI Holy Trinity)
     
    Last edited: Dec 16, 2018
    NAZ2222 likes this.
  10. OnnA

    OnnA Ancient Guru

    Messages:
    17,851
    Likes Received:
    6,741
    GPU:
    TiTan RTX Ampere UV
    Ok, i will....
     
    jimpsar likes this.

  11. OnnA

    OnnA Ancient Guru

    Messages:
    17,851
    Likes Received:
    6,741
    GPU:
    TiTan RTX Ampere UV
    1775MHz/1200HBM2 at 1.131v/1050v
    Almost 28k (27.932) -> deleted result, was too low ;)
    I was testing new Adrenalin 2019.

    So i need 1.106mV for 1750MHz
    Note:
    (Every chip is a little different, my GPU is selected - it's the best from 5 tested Vega XTX, shop was kind enough to test those especially for me :rolleyes:)

    Every Vega 64 (and even V56) can operate with 1.100v easily.
    One can hit 1680 and another 1700MHz, most of the XTX can hit 1730-1750MHz with such low v.

    [​IMG]
     
    Last edited: Dec 16, 2018
  12. OnnA

    OnnA Ancient Guru

    Messages:
    17,851
    Likes Received:
    6,741
    GPU:
    TiTan RTX Ampere UV
    Op Updated (post no.2)

    How to set & test Lower P-states like. P1->P4
     
  13. slicer

    slicer Member Guru

    Messages:
    140
    Likes Received:
    51
    GPU:
    Sapphire Vega 64
    Thank You!
     
    OnnA likes this.
  14. Kuski

    Kuski Guest

    Messages:
    14
    Likes Received:
    1
    GPU:
    Asus Strix vega 64
    Hello. I don't know if this is the right place to ask, but I need to know what should I do with my strix vega 64? I'm not really getting the performance out of it as I expected. Here is screenshot as its running a stress test. (Pretty heavily undervolted, slight bump in memory clock and custom fan profile). I changed the thermal pad to thermal grizzly and the paste to to noctua nt h1 (I used a lot). The biggest problem here is the gpu hotspot. Its so damn hot. Ok so here is the screenshot: [​IMG]
     
  15. OnnA

    OnnA Ancient Guru

    Messages:
    17,851
    Likes Received:
    6,741
    GPU:
    TiTan RTX Ampere UV
    Every one has hot Hot-Spot ;)
    If not exceeding 85deg then it's OK IMhO (some say it will be ok at even 90deg)
    My Goes highest (in summer) to 75Deg, while the rest GPU/HBM2 stays at ~60-62deg
    Hotspot usually is at +10-15deg mark, so 60deg GPU and 70deg Hotspot is fine.
     

  16. LocoDiceGR

    LocoDiceGR Ancient Guru

    Messages:
    2,860
    Likes Received:
    1,073
    GPU:
    Gigabyte 3060 Ti
    @OnnA I need your help, my friend wants the HEVC exe or something to make it work..is this possible?
     
  17. Kuski

    Kuski Guest

    Messages:
    14
    Likes Received:
    1
    GPU:
    Asus Strix vega 64
    It is actually exceeding 90 degrees even when undervolted. Right when I open a game even if gpu core is still warming up from 50 degrees the hotspot is already 80+. I have tried so many different overclocks, but I think it is impossible to overclock this card at all.
     
  18. OnnA

    OnnA Ancient Guru

    Messages:
    17,851
    Likes Received:
    6,741
    GPU:
    TiTan RTX Ampere UV
    You need to test it, Play game, monitor Temps with HWinfo64
    Then if GPU works OK (despite those Temps) then maby it is OK.
    Every GPU is different.

    @JonasBeckman can you elaborate on?
     
    Last edited: Dec 16, 2018
  19. OnnA

    OnnA Ancient Guru

    Messages:
    17,851
    Likes Received:
    6,741
    GPU:
    TiTan RTX Ampere UV
    What's this?
     
  20. JonasBeckman

    JonasBeckman Ancient Guru

    Messages:
    17,564
    Likes Received:
    2,961
    GPU:
    XFX 7900XTX M'310
    The hotspot?

    I know AMD is being vague about what it really is and there's been some suggestions from the high-bandwidth memory interposer to the voltage doublers with heat cameras and monitoring tools trying to figure that one out, threshold seems to be 105 degrees Celsius before the GPU throttles and AMD also requested monitoring being off by default in program such as GPU-Z since it's often misunderstood.

    I mainly know what I've been reading up on about it but there's nothing truly conclusive AFAIK though some of the info is probably pretty close to how it works but AMD isn't saying much and that's certainly not helping.

    It does tend to run hot though if it's even temperature at all and not a sensor that just appears to look like temperature since some tests have it shoot up immediately instead of gradually when measured so it's weird and yeah different GPU's produce different results from 80 to 90 or even 100 or up to 115 degrees which that would be a issue if the GPU starts trying to throttle down if this hits 105 degrees C and then some cards run a bit above that value by default and no further overclocking or modifications.
     
Thread Status:
Not open for further replies.

Share This Page