Direct3D feature levels discussion

Discussion in 'Rendering Technology and APIs' started by DmitryKo, Feb 20, 2015.

  1. CarstenS

    Legend Subscriber

    Joined:
    May 31, 2002
    Messages:
    5,800
    Likes Received:
    3,920
    Location:
    Germany
    That makes even more sense, because you have precedence.

    edit: If you want to entertain that train of thought in the first place and do not dimiss it as conspiracy theory outright, which I would totally understand.
     
    Krteq likes this.
  2. Jawed

    Legend

    Joined:
    Oct 2, 2004
    Messages:
    11,714
    Likes Received:
    2,135
    Location:
    London
    SM6.6 looks pretty interesting. Is this mainly about XBox Series, with forwards compatibility over time for PC?
     
  3. CarstenS

    Legend Subscriber

    Joined:
    May 31, 2002
    Messages:
    5,800
    Likes Received:
    3,920
    Location:
    Germany
    #1163 CarstenS, Jul 21, 2021
    Last edited: Jul 23, 2021
    BRiT and Malo like this.
  4. DegustatoR

    Veteran

    Joined:
    Mar 12, 2002
    Messages:
    3,242
    Likes Received:
    3,403
    Latest version of the D3D12CheckFeatureSupport tool has stopped reporting anything after SharedResourceCompatibilityTier : D3D12_SHARED_RESOURCE_COMPATIBILITY_TIER_0 (0) for some reason...

    The output looks like this now:

    Code:
    Direct3D 12 feature checker (April 2021) by DmitryKo (x64)
    https://forum.beyond3d.com/posts/1840641/
    
    Windows 6.2 version 2009 (build 9200.1288 vb_release) x64
    
    ADAPTER 0
    "NVIDIA GeForce RTX 3080"
    VEN_10DE, DEV_2206, SUBSYS_87AC1043, REV_A1
    Dedicated video memory : 10078.0 MB (10567548928 bytes)
    Total video memory : 42802.8 MB (44881985536 bytes)
    Video driver version : 30.0.14.9613
    WDDM version : KMT_DRIVERVERSION_WDDM_2_7 (2700)
    Virtual memory model : GPUMMU
    Hardware-accelerated scheduler : Enabled
    GraphicsPreemptionGranularity : DXGI_GRAPHICS_PREEMPTION_PIXEL_BOUNDARY (3)
    ComputePreemptionGranularity : DXGI_COMPUTE_PREEMPTION_DISPATCH_BOUNDARY (1)
    Maximum feature level : D3D_FEATURE_LEVEL_12_1 (0xc100)
    DoublePrecisionFloatShaderOps : 1
    OutputMergerLogicOp : 1
    MinPrecisionSupport : D3D12_SHADER_MIN_PRECISION_SUPPORT_16_BIT (2) (0b0000'0010)
    TiledResourcesTier : D3D12_TILED_RESOURCES_TIER_3 (3)
    ResourceBindingTier : D3D12_RESOURCE_BINDING_TIER_3 (3)
    PSSpecifiedStencilRefSupported : 0
    TypedUAVLoadAdditionalFormats : 1
    ROVsSupported : 1
    ConservativeRasterizationTier : D3D12_CONSERVATIVE_RASTERIZATION_TIER_3 (3)
    StandardSwizzle64KBSupported : 0
    CrossNodeSharingTier : D3D12_CROSS_NODE_SHARING_TIER_NOT_SUPPORTED (0)
    CrossAdapterRowMajorTextureSupported : 0
    VPAndRTArrayIndexFromAnyShaderFeedingRasterizerSupportedWithoutGSEmulation : 1
    ResourceHeapTier : D3D12_RESOURCE_HEAP_TIER_2 (2)
    MaxGPUVirtualAddressBitsPerResource : 40
    MaxGPUVirtualAddressBitsPerProcess : 40
    Adapter Node 0:     TileBasedRenderer: 0, UMA: 0, CacheCoherentUMA: 0DisplayableTexture.SharedResourceCompatibilityTier : D3D12_SHARED_RESOURCE_COMPATIBILITY_TIER_0 (0)
    
     
  5. Osamar

    Newcomer

    Joined:
    Sep 19, 2006
    Messages:
    231
    Likes Received:
    43
    Location:
    40,00ºN - 00,00ºE
    Minimum video driver version for AMD hardware video encoding "In development – ETA Q2 ‘2022" :huh:
     
  6. Lurkmass

    Regular

    Joined:
    Mar 3, 2020
    Messages:
    565
    Likes Received:
    711
    D3D12 "enhanced barriers" is virtually a clone of Vulkan's barrier model. Vulkan just got rid of render passes recently to be more like D3D12. Wow, what a turnaround ...

    Basically the latest revisions between D3D12 and Vulkan aren't all that different anymore in terms of verbosity.
     
  7. Alessio1989

    Regular

    Joined:
    Jun 6, 2015
    Messages:
    614
    Likes Received:
    321
    revisioned barriers and more C++ like HLSL are bigger deal than a video encoding api upon d3d12 interfaces to be honest.
     
    #1171 Alessio1989, Dec 10, 2021
    Last edited: Dec 11, 2021
    Krteq likes this.
  8. Osamar

    Newcomer

    Joined:
    Sep 19, 2006
    Messages:
    231
    Likes Received:
    43
    Location:
    40,00ºN - 00,00ºE
    I agree, but 4 to 7 months for adding video encoding to a driver has surprised me.
     
  9. Lurkmass

    Regular

    Joined:
    Mar 3, 2020
    Messages:
    565
    Likes Received:
    711
    C++ templates for shaders might not be as good of an idea as we once thought ...



    A part of the reason why Tensorflow has ridiculously long build times is that they abuse templates and the Visual Studio team at Microsoft specifically tried to improve their case. Templates could potentially blow up shader compilation times ...
     
    Jawed, pharma, CeeGee and 2 others like this.
  10. Alessio1989

    Regular

    Joined:
    Jun 6, 2015
    Messages:
    614
    Likes Received:
    321
    you can abuse for loops too if you are dumb enough :V
     
  11. DegustatoR

    Veteran

    Joined:
    Mar 12, 2002
    Messages:
    3,242
    Likes Received:
    3,403
  12. DegustatoR

    Veteran

    Joined:
    Mar 12, 2002
    Messages:
    3,242
    Likes Received:
    3,403
  13. Remij

    Regular

    Joined:
    May 3, 2008
    Messages:
    678
    Likes Received:
    1,258
  14. DegustatoR

    Veteran

    Joined:
    Mar 12, 2002
    Messages:
    3,242
    Likes Received:
    3,403
    I don't think anything was disclosed so far.
    Considering that it's coming around the time of Lovelace and RDNA3 launches it can be either another minor update or something more serious.
     
    BRiT and Remij like this.
Loading...

Share This Page

  • About Us

    Beyond3D has been around for over a decade and prides itself on being the best place on the web for in-depth, technically-driven discussion and analysis of 3D graphics hardware. If you love pixels and transistors, you've come to the right place!

    Beyond3D is proudly published by GPU Tools Ltd.
Loading...