RX 580 4GB Benchmark Scores: Is this OK/Normal??

Discussion in 'Videocards - AMD Radeon' started by jortego128, Oct 11, 2017.

  1. jortego128

    jortego128 Guest

    Messages:
    107
    Likes Received:
    15
    GPU:
    AMD RX 580 4GB
    I recently purchased a very slightly used (at least its physically flawless) ASUS Dual Fan RX 580 OC 4GB card from Ebay. Upon popping it in my PC found that its BIOS had been modded to GPU/Mem 1150MHz/2000MHz for mining-- this was not advertised in the Ebay listing-- the card was underperforming in gaming so I went through the steps and obtained a default BIOS from TechPowerup.com, flashed the GPU, all went well, and performance did increase.

    However, I have 2 issues/questions-- in Firestrike 1.1 and Timespy, before I run a benchmark, the opening menu where you select your bench recognizes the GPU as a RX 580, but after I run a benchmark, I get a validation warning on both FS and TS saying my GPU and my driver version are not recognized. Im using Relive 17.9.2 I believe. I tried using DDU to clean the drivers and have re-installed different versions, but I still get the same validation warning-- is this just because the ASUS Dual 4GB may have not been submitted to the 3dmark database yet, or something more sinister?

    Second question is performance-- Im running a Ryzen 1700@3.6 GHz, 8G 2666MHz RAM, SSD, and default clocks on the RX 580. I dont have the exact scores handy now, but Im basically getting the scores below

    [​IMG]



    Firestrike:
    Graphics (dont remember exactly, will update later today)
    Overall - 11400

    While I dont think these scores are horrible, they are falling about 10% behind what Im reading on average for a 580. That said, almost all 580 reviews I have found are using an 8GB version, which also has 2000MHz mem clock vs. 1750MHz mem clock of all 4GB versions. I havent yet tried to OC the memory to 2000 to see how it affects the run. Does anyone know the average performance delta on FS and TS with the different mem sizes and speeds?
     
    Last edited: Oct 11, 2017
  2. jortego128

    jortego128 Guest

    Messages:
    107
    Likes Received:
    15
    GPU:
    AMD RX 580 4GB
    Over 50 views and nobody knows/has a 580??
     
  3. OnnA

    OnnA Ancient Guru

    Messages:
    17,963
    Likes Received:
    6,823
    GPU:
    TiTan RTX Ampere UV
    I think is a good score :D
    Some Games have Ultra Textures for 4K ->
    Try it on High, to see if this helps, also Shadows on High.

    Based on the score (My FuryX has 5230) is good one and you can Play any game :p
    In all games try to find best Quality vs Performance scenarios in Options/Settings

    Ne01 out.
     
  4. Valerys

    Valerys Master Guru

    Messages:
    395
    Likes Received:
    18
    GPU:
    Gigabyte RTX2080S
    My (slightly) OCed R290 gets almost the same graphics score if it helps.
    https://www.3dmark.com/spy/2489309

    Since RX580 has only a 256bit bus, changes in VRAM speeds are more likely to affect the performance.
     
    Last edited: Oct 12, 2017

  5. OnnA

    OnnA Ancient Guru

    Messages:
    17,963
    Likes Received:
    6,823
    GPU:
    TiTan RTX Ampere UV
    PPL please don't rely on benchmarks tooo much :D it's for fun only (I like GFX in 3Dmarks, Yes i have all of'em 99,2000 etc.)
    just play you favorite games and see how it runs.

     
  6. jortego128

    jortego128 Guest

    Messages:
    107
    Likes Received:
    15
    GPU:
    AMD RX 580 4GB
    Thanks-- with a bit of undervolting and raising the power limit a bit Im hitting 4500+ overall score. I now think this is OK for a 580, knowing it hangs with a 290 is icing on the cake!
     
  7. jortego128

    jortego128 Guest

    Messages:
    107
    Likes Received:
    15
    GPU:
    AMD RX 580 4GB
    Thanks for the FuryX reference. I do think I have the bios issues ironed out now and think my card is doing what its supposed to.
     

Share This Page