Fix validation errors

AnKi/Gr/Common.h

@@ -520,6 +520,8 @@ enum class TextureUsageBit : U32
 				| kTransferDestination | kGenerateMipmaps,
 	kAll = kAllRead | kAllWrite,
 	kAllShaderResource = kAllSampled | kAllStorage,
+	kAllSrv = (kAllSampled | kAllStorage) & kAllRead,
+	kAllUav = kAllStorage & kAllWrite,
 };
 ANKI_ENUM_ALLOW_NUMERIC_OPERATIONS(TextureUsageBit)
 

AnKi/Gr/D3D/D3DGrManager.cpp

@@ -29,7 +29,7 @@
 
 // Use the Agility SDK
 extern "C" {
-__declspec(dllexport) extern const UINT D3D12SDKVersion = 613; // Number taken from the download page
+__declspec(dllexport) extern const UINT D3D12SDKVersion = 614; // Number taken from the download page
 __declspec(dllexport) extern const char* D3D12SDKPath = ".\\"; // The D3D12Core.dll should be in the same dir as the .exe
 }
 
@@ -41,6 +41,7 @@ BoolCVar g_vsyncCVar(CVarSubsystem::kGr, "Vsync", false, "Enable or not vsync");
 BoolCVar g_debugMarkersCVar(CVarSubsystem::kGr, "DebugMarkers", false, "Enable or not debug markers");
 BoolCVar g_meshShadersCVar(CVarSubsystem::kGr, "MeshShaders", false, "Enable or not mesh shaders");
 static NumericCVar<U8> g_deviceCVar(CVarSubsystem::kGr, "Device", 0, 0, 16, "Choose an available device. Devices are sorted by performance");
+static BoolCVar g_rayTracingCVar(CVarSubsystem::kGr, "RayTracing", false, "Try enabling ray tracing");
 
 static LONG NTAPI vexHandler(PEXCEPTION_POINTERS exceptionInfo)
 {
@@ -407,7 +408,7 @@ Error GrManagerImpl::initInternal(const GrManagerInitInfo& init)
 
 	// Create device
 	ComPtr<ID3D12Device> dev;
-	ANKI_D3D_CHECK(D3D12CreateDevice(adapters[chosenPhysDevIdx].m_adapter.Get(), D3D_FEATURE_LEVEL_12_1, IID_PPV_ARGS(&dev)));
+	ANKI_D3D_CHECK(D3D12CreateDevice(adapters[chosenPhysDevIdx].m_adapter.Get(), D3D_FEATURE_LEVEL_12_2, IID_PPV_ARGS(&dev)));
 	ANKI_D3D_CHECK(dev->QueryInterface(IID_PPV_ARGS(&m_device)));
 
 	if(g_validationCVar.get())
@@ -426,17 +427,6 @@ Error GrManagerImpl::initInternal(const GrManagerInitInfo& init)
 		}
 	}
 
-	// Capabilities
-	{
-		D3D12_FEATURE_DATA_D3D12_OPTIONS16 options16;
-		ANKI_D3D_CHECK(m_device->CheckFeatureSupport(D3D12_FEATURE_D3D12_OPTIONS16, &options16, sizeof(options16)));
-		m_d3dCapabilities.m_rebar = options16.GPUUploadHeapSupported;
-		if(!m_d3dCapabilities.m_rebar)
-		{
-			ANKI_D3D_LOGW("ReBAR not supported");
-		}
-	}
-
 	// Create queues
 	{
 		D3D12_COMMAND_QUEUE_DESC queueDesc = {};
@@ -449,6 +439,43 @@ Error GrManagerImpl::initInternal(const GrManagerInitInfo& init)
 		ANKI_D3D_CHECK(m_queues[GpuQueueType::kGeneral]->GetTimestampFrequency(&m_timestampFrequency));
 	}
 
+	// Set device capabilities (taken from mesa's dozen driver)
+	{
+		D3D12_FEATURE_DATA_D3D12_OPTIONS16 options16;
+		ANKI_D3D_CHECK(m_device->CheckFeatureSupport(D3D12_FEATURE_D3D12_OPTIONS16, &options16, sizeof(options16)));
+		D3D12_FEATURE_DATA_ARCHITECTURE architecture = {.NodeIndex = 0};
+		ANKI_D3D_CHECK(m_device->CheckFeatureSupport(D3D12_FEATURE_ARCHITECTURE, &architecture, sizeof(architecture)));
+
+		m_d3dCapabilities.m_rebar = options16.GPUUploadHeapSupported;
+		if(!m_d3dCapabilities.m_rebar)
+		{
+			ANKI_D3D_LOGW("ReBAR not supported");
+		}
+
+		m_capabilities.m_uniformBufferBindOffsetAlignment = D3D12_CONSTANT_BUFFER_DATA_PLACEMENT_ALIGNMENT;
+		m_capabilities.m_uniformBufferMaxRange = D3D12_REQ_CONSTANT_BUFFER_ELEMENT_COUNT * D3D12_STANDARD_VECTOR_SIZE * sizeof(F32);
+		m_capabilities.m_storageBufferBindOffsetAlignment = D3D12_RAW_UAV_SRV_BYTE_ALIGNMENT;
+		m_capabilities.m_storageBufferMaxRange = 1 << D3D12_REQ_BUFFER_RESOURCE_TEXEL_COUNT_2_TO_EXP;
+		m_capabilities.m_texelBufferBindOffsetAlignment = 32;
+		m_capabilities.m_textureBufferMaxRange = kMaxU32; // ?
+		m_capabilities.m_pushConstantsSize = kMaxPushConstantSize;
+		m_capabilities.m_computeSharedMemorySize = D3D12_CS_TGSM_REGISTER_COUNT * sizeof(F32);
+		m_capabilities.m_accelerationStructureBuildScratchOffsetAlignment = 32; // ?
+		m_capabilities.m_sbtRecordAlignment = 32; // ?
+		m_capabilities.m_maxDrawIndirectCount = kMaxU32;
+		m_capabilities.m_discreteGpu = !architecture.UMA;
+		m_capabilities.m_majorApiVersion = 12;
+		m_capabilities.m_rayTracingEnabled = g_rayTracingCVar.get();
+		m_capabilities.m_64bitAtomics = true;
+		m_capabilities.m_vrs = true;
+		m_capabilities.m_samplingFilterMinMax = true;
+		m_capabilities.m_unalignedBbpTextureFormats = false;
+		m_capabilities.m_dlss = false;
+		m_capabilities.m_meshShaders = true;
+		m_capabilities.m_pipelineQuery = true;
+		m_capabilities.m_barycentrics = true;
+	}
+
 	// Other systems
 	DescriptorFactory::allocateSingleton();
 	ANKI_CHECK(DescriptorFactory::getSingleton().init());
@@ -504,11 +531,14 @@ void GrManagerImpl::destroy()
 
 void GrManagerImpl::waitAllQueues()
 {
-	for(GpuQueueType queueType : EnumIterable<GpuQueueType>())
+	if(FenceFactory::isAllocated())
 	{
-		MicroFencePtr fence = FenceFactory::getSingleton().newInstance();
-		fence->gpuSignal(queueType);
-		fence->clientWait(kMaxSecond);
+		for(GpuQueueType queueType : EnumIterable<GpuQueueType>())
+		{
+			MicroFencePtr fence = FenceFactory::getSingleton().newInstance();
+			fence->gpuSignal(queueType);
+			fence->clientWait(kMaxSecond);
+		}
 	}
 }
 

AnKi/Gr/D3D/D3DTexture.cpp

@@ -131,7 +131,7 @@ Error TextureImpl::initInternal(ID3D12Resource* external, const TextureInitInfo&
 		desc.SampleDesc.Count = 1;
 		desc.SampleDesc.Quality = 0;
 		desc.Layout = D3D12_TEXTURE_LAYOUT_UNKNOWN;
-		desc.Flags = D3D12_RESOURCE_FLAG_ALLOW_SIMULTANEOUS_ACCESS;
+		desc.Flags = {};
 
 		if(!!(m_usage & TextureUsageBit::kAllFramebuffer) && m_aspect == DepthStencilAspectBit::kNone)
 		{
@@ -257,9 +257,49 @@ void TextureImpl::initView(const TextureSubresourceDescriptor& subresource, Text
 	{
 		const TextureView tview(this, subresource);
 
+		ANKI_ASSERT(tview.getSubresource().m_depthStencilAspect != DepthStencilAspectBit::kDepthStencil && "Can only create a single plane SRV");
+
 		D3D12_SHADER_RESOURCE_VIEW_DESC desc = {};
 
-		desc.Format = DXGI_FORMAT(m_format);
+		if(!m_aspect)
+		{
+			desc.Format = DXGI_FORMAT(m_format);
+		}
+		else
+		{
+			// D3D doesn't like depth formats as SRVs
+			switch(m_format)
+			{
+			case Format::kD16_Unorm:
+				desc.Format = DXGI_FORMAT(Format::kR16_Unorm);
+				break;
+			case Format::kD32_Sfloat:
+				desc.Format = DXGI_FORMAT(Format::kR32_Sfloat);
+				break;
+			case Format::kD24_Unorm_S8_Uint:
+				if(tview.getSubresource().m_depthStencilAspect == DepthStencilAspectBit::kDepth)
+				{
+					desc.Format = DXGI_FORMAT_R24_UNORM_X8_TYPELESS;
+				}
+				else
+				{
+					desc.Format = DXGI_FORMAT_X24_TYPELESS_G8_UINT;
+				}
+				break;
+			case Format::kD32_Sfloat_S8_Uint:
+				if(tview.getSubresource().m_depthStencilAspect == DepthStencilAspectBit::kDepth)
+				{
+					desc.Format = DXGI_FORMAT_R32_FLOAT_X8X24_TYPELESS;
+				}
+				else
+				{
+					desc.Format = DXGI_FORMAT_R32G8X24_TYPELESS;
+				}
+				break;
+			default:
+				ANKI_ASSERT(0);
+			}
+		}
 
 		const U32 faceCount = textureTypeIsCube(m_texType) ? 6 : 1;
 		const U32 surfaceCount = faceCount * m_layerCount;
@@ -501,7 +541,16 @@ D3D12_TEXTURE_BARRIER TextureImpl::computeBarrierInfo(TextureUsageBit before, Te
 		barrier.Subresources.NumMipLevels = 1;
 		barrier.Subresources.FirstArraySlice = subresource.m_layer * faceCount + subresource.m_face;
 		barrier.Subresources.NumArraySlices = 1;
-		barrier.Subresources.FirstPlane = !!(subresource.m_depthStencilAspect & DepthStencilAspectBit::kDepth) ? 0 : 1;
+
+		if(!!(subresource.m_depthStencilAspect & DepthStencilAspectBit::kDepth) || subresource.m_depthStencilAspect == DepthStencilAspectBit::kNone)
+		{
+			barrier.Subresources.FirstPlane = 0;
+		}
+		else
+		{
+			barrier.Subresources.FirstPlane = 1;
+		}
+
 		barrier.Subresources.NumPlanes = (subresource.m_depthStencilAspect == DepthStencilAspectBit::kDepthStencil) ? 2 : 1;
 	}
 
@@ -638,36 +687,46 @@ D3D12_BARRIER_LAYOUT TextureImpl::computeLayout(TextureUsageBit usage) const
 	}
 	else if(depthStencil)
 	{
-		if(!(usage & ~(TextureUsageBit::kAllSampled | TextureUsageBit::kFramebufferRead)))
+		if((usage & TextureUsageBit::kAllSampled) == usage)
 		{
-			// Only depth tests and sampled
+			// Only sampled
+			out = D3D12_BARRIER_LAYOUT_SHADER_RESOURCE;
+		}
+		else if((usage & TextureUsageBit::kFramebufferRead) == usage)
+		{
+			// Only FB read
 			out = D3D12_BARRIER_LAYOUT_DEPTH_STENCIL_READ;
 		}
+		else if((usage & (TextureUsageBit::kAllSampled | TextureUsageBit::kFramebufferRead)) == usage)
+		{
+			// Only depth tests and sampled
+			out = D3D12_BARRIER_LAYOUT_COMMON;
+		}
 		else
 		{
 			// Only attachment write, the rest (eg transfer) are not supported for now
-			ANKI_ASSERT(usage == TextureUsageBit::kFramebufferWrite || usage == TextureUsageBit::kAllFramebuffer);
+			ANKI_ASSERT(usage == TextureUsageBit::kAllFramebuffer || usage == TextureUsageBit::kFramebufferWrite);
 			out = D3D12_BARRIER_LAYOUT_DEPTH_STENCIL_WRITE;
 		}
 	}
-	else if(!(usage & ~TextureUsageBit::kAllFramebuffer))
+	else if((usage & TextureUsageBit::kAllFramebuffer) == usage)
 	{
 		// Color attachment
 		out = D3D12_BARRIER_LAYOUT_RENDER_TARGET;
 	}
-	else if(!(usage & ~TextureUsageBit::kFramebufferShadingRate))
+	else if((usage & TextureUsageBit::kFramebufferShadingRate) == usage)
 	{
 		// SRI
 		out = D3D12_BARRIER_LAYOUT_SHADING_RATE_SOURCE;
 	}
-	else if(!(usage & ~TextureUsageBit::kAllStorage))
+	else if((usage & TextureUsageBit::kAllUav) == usage)
 	{
-		// Only image load/store
+		// UAV
 		out = D3D12_BARRIER_LAYOUT_UNORDERED_ACCESS;
 	}
-	else if(!(usage & ~TextureUsageBit::kAllSampled))
+	else if((usage & TextureUsageBit::kAllSrv) == usage)
 	{
-		// Only sampled
+		// SRV
 		out = D3D12_BARRIER_LAYOUT_SHADER_RESOURCE;
 	}
 	else if(usage == TextureUsageBit::kGenerateMipmaps)

AnKi/Renderer/Tonemapping.cpp

@@ -38,15 +38,16 @@ Error Tonemapping::initInternal()
 	const TextureInitInfo texinit = getRenderer().create2DRenderTargetInitInfo(1, 1, Format::kR16G16_Sfloat, usage, "ExposureAndAvgLum1x1");
 	ClearValue clearValue;
 	clearValue.m_colorf = {0.5f, 0.5f, 0.5f, 0.5f};
-	m_exposureAndAvgLuminance1x1 = getRenderer().createAndClearRenderTarget(texinit, TextureUsageBit::kAllStorage, clearValue);
+	m_exposureAndAvgLuminance1x1 = getRenderer().createAndClearRenderTarget(texinit, TextureUsageBit::kStorageComputeRead, clearValue);
 
 	return Error::kNone;
 }
 
 void Tonemapping::importRenderTargets(RenderingContext& ctx)
 {
 	// Just import it. It will not be used in resource tracking
-	m_runCtx.m_exposureLuminanceHandle = ctx.m_renderGraphDescr.importRenderTarget(m_exposureAndAvgLuminance1x1.get(), TextureUsageBit::kAllStorage);
+	m_runCtx.m_exposureLuminanceHandle =
+		ctx.m_renderGraphDescr.importRenderTarget(m_exposureAndAvgLuminance1x1.get(), TextureUsageBit::kStorageComputeRead);
 }
 
 void Tonemapping::populateRenderGraph(RenderingContext& ctx)

ThirdParty/AgilitySdk/README.md

@@ -13,40 +13,8 @@ The included licenses apply to the following files:
 
 ## Changelog
 
-### Version 1.613.3
-
-- Same as 1.613.2, with minor updates shown at the end of this list:
-- Work Graphs
-- Generic Programs in State objects
-- Shader Model 6.8
-  - Work Graphs support
-  - Start Vertex/Instance Location
-  - Wave Size Range
-  - Expanded Comparison Sampling
-- GPU Upload Heaps (requires preview or future OS)
-- Incrementing Constants in ExecuteIndirect
-- Minor updates for .3:
-   - Honor root signature associations added to export in collections when the export is imported into generic program in an executable state object.
-   - Propagate root signatures to exports that have bindings and no root signatures within the scope of a generic program, like what happens with PSOs
-   - Other ganeric programs releated bug fixes, additional validation and optimizations
-   - For state objectgs, minor fixes to subobject association logic in edge cases involving collections to more precisely match spec wording
-
-### Version 1.613.2
-
-- Same as 1.613.1, with minor updates shown at the end of this list:
-- Work Graphs
-- Generic Programs in State objects
-- Shader Model 6.8
-  - Work Graphs support
-  - Start Vertex/Instance Location
-  - Wave Size Range
-  - Expanded Comparison Sampling
-- GPU Upload Heaps (requires preview or future OS)
-- Incrementing Constants in ExecuteIndirect
-- Minor updates for .2:
-   - Close validation hole: Prevent use of features in incompatible state object types, such as defining generic program components (like rast state) in a raytracing pipeline.
-   - Debug layer fixes around lifetime management of program identifiers (e.g. the handles for the new generic programs)
-   - ExecuteIndirect validation was too strict: buffer size only needs to be big enough for the number of commands * stride, _minus the unused space at end of last command_
+### Version 1.614.0
+- Enabled R9G9B9E5_SHAREDEXP format for Render Target and Unordered Access Views
 
 ### Version 1.613.1
 

ThirdParty/AgilitySdk/include/d3d12.h

@@ -1102,7 +1102,7 @@ extern "C"{
 
 #define	D3D12_PIXEL_ADDRESS_RANGE_BIT_COUNT	( 15 )
 
-#define	D3D12_PREVIEW_SDK_VERSION	( 713 )
+#define	D3D12_PREVIEW_SDK_VERSION	( 714 )
 
 #define	D3D12_PRE_SCISSOR_PIXEL_ADDRESS_RANGE_BIT_COUNT	( 16 )
 
@@ -1235,7 +1235,7 @@ extern "C"{
 
 #define	D3D12_RS_SET_SHADING_RATE_COMBINER_COUNT	( 2 )
 
-#define	D3D12_SDK_VERSION	( 613 )
+#define	D3D12_SDK_VERSION	( 614 )
 
 #define	D3D12_SHADER_IDENTIFIER_SIZE_IN_BYTES	( 32 )
 

ThirdParty/AgilitySdk/include/d3d12.idl

@@ -293,7 +293,7 @@ const UINT D3D12_OS_RESERVED_REGISTER_SPACE_VALUES_END = 0xffffffff;
 const UINT D3D12_OS_RESERVED_REGISTER_SPACE_VALUES_START = 0xfffffff8;
 const UINT D3D12_PACKED_TILE = 0xffffffff;
 const UINT D3D12_PIXEL_ADDRESS_RANGE_BIT_COUNT = 15;
-const UINT D3D12_PREVIEW_SDK_VERSION = 713;
+const UINT D3D12_PREVIEW_SDK_VERSION = 714;
 const UINT D3D12_PRE_SCISSOR_PIXEL_ADDRESS_RANGE_BIT_COUNT = 16;
 const UINT D3D12_PS_CS_UAV_REGISTER_COMPONENTS = 1;
 const UINT D3D12_PS_CS_UAV_REGISTER_COUNT = 8;
@@ -361,7 +361,7 @@ const UINT D3D12_REQ_TEXTURECUBE_DIMENSION = 16384;
 const UINT D3D12_RESINFO_INSTRUCTION_MISSING_COMPONENT_RETVAL = 0;
 const UINT D3D12_RESOURCE_BARRIER_ALL_SUBRESOURCES = 0xffffffff;
 const UINT D3D12_RS_SET_SHADING_RATE_COMBINER_COUNT = 2;
-const UINT D3D12_SDK_VERSION = 613;
+const UINT D3D12_SDK_VERSION = 614;
 const UINT D3D12_SHADER_IDENTIFIER_SIZE_IN_BYTES = 32;
 const UINT D3D12_SHADER_MAJOR_VERSION = 5;
 const UINT D3D12_SHADER_MAX_INSTANCES = 65535;
@@ -1030,7 +1030,7 @@ typedef enum D3D12_FEATURE
     D3D12_FEATURE_PREDICATION                           = 50,
     D3D12_FEATURE_PLACED_RESOURCE_SUPPORT_INFO          = 51,
     D3D12_FEATURE_HARDWARE_COPY                         = 52,
-    D3D12_FEATURE_D3D12_OPTIONS21                       = 53
+    D3D12_FEATURE_D3D12_OPTIONS21                       = 53,
 } D3D12_FEATURE;
 
 typedef enum D3D12_SHADER_MIN_PRECISION_SUPPORT
@@ -1621,6 +1621,7 @@ typedef struct D3D12_FEATURE_DATA_HARDWARE_COPY
     [annotation("_Out_")] BOOL Supported;
 } D3D12_FEATURE_DATA_HARDWARE_COPY;
 
+
 typedef struct D3D12_RESOURCE_ALLOCATION_INFO
 {
     UINT64 SizeInBytes;
@@ -1779,6 +1780,7 @@ typedef struct D3D12_RESOURCE_DESC1
 } D3D12_RESOURCE_DESC1;
 
 
+
 typedef struct D3D12_DEPTH_STENCIL_VALUE
 {
     FLOAT Depth;
@@ -5430,6 +5432,7 @@ interface ID3D12Resource2
 }
 
 
+
 [uuid(572F7389-2168-49E3-9693-D6DF5871BF6D), object, local, pointer_default(unique)]
 interface ID3D12Heap1
     : ID3D12Heap
@@ -6063,6 +6066,7 @@ interface ID3D12Device14 : ID3D12Device13
         [out, iid_is(riid), annotation("_COM_Outptr_")] void** ppvRootSignature);
 };
 
+
 [uuid(bc66d368-7373-4943-8757-fc87dc79e476), object, local, pointer_default(unique)]
 interface ID3D12VirtualizationGuestDevice
     : IUnknown
@@ -6089,7 +6093,6 @@ interface ID3D12Tools
 }
 
 
-
 typedef struct D3D12_SUBRESOURCE_DATA
 {
     const void* pData;