v1.1.0

Author: kurbeco

.gitattributes

@@ -0,0 +1,8 @@
+libs/Cauldron/libs/DXC/bin/dxcompiler.dll filter=lfs diff=lfs merge=lfs -text
+libs/Cauldron/libs/DXC/bin/dxc.exe filter=lfs diff=lfs merge=lfs -text
+libs/Cauldron/libs/DXC/bin/dxil.dll filter=lfs diff=lfs merge=lfs -text
+libs/Cauldron/libs/DXC/bin/x64/dxcompiler.dll filter=lfs diff=lfs merge=lfs -text
+libs/Cauldron/libs/DXC/bin/x64/dxc.exe filter=lfs diff=lfs merge=lfs -text
+libs/Cauldron/libs/DXC/bin/x64/dxil.dll filter=lfs diff=lfs merge=lfs -text
+libs/Cauldron/libs/DXCwg/*.exe filter=lfs diff=lfs merge=lfs -text
+libs/Cauldron/libs/DXCwg/*.dll filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -355,4 +355,8 @@ ASALocalRun/
 healthchecksdb
 
 # Backup folder for Package Reference Convert tool in Visual Studio 2017
-MigrationBackup/
+MigrationBackup/
+/libs/Cauldron/libs/DXCwg/*.exe
+/libs/Cauldron/libs/DXCwg/*.dll
+/libs/Cauldron/libs/DXCwg/LICENSE*.txt
+/libs/Cauldron/libs/DXCwg/LICENCE*.txt

.gitmodules

@@ -1,7 +1,7 @@
 [submodule "media/Cauldron-Media"]
 	path = media/Cauldron-Media
-	url = ../Cauldron-Media.git
+	url = https://github.com/GPUOpen-LibrariesAndSDKs/Cauldron-Media.git
 [submodule "libs/Cauldron"]
-	path = libs/Cauldron
-	url = ../Cauldron.git
-	branch = WorkGraphComputeRasterizer
+        path = libs/Cauldron
+        url = ../Cauldron.git
+        branch = WorkGraphComputeRasterizer

CMakeLists.txt

@@ -46,7 +46,7 @@ endforeach( OUTPUTCONFIG CMAKE_CONFIGURATION_TYPES )
 add_compile_options(/MP)
 
 if(GFX_API_WG)
-    add_definitions(-DENABLE_EXPERIMENTAL_WORKGRAPHS)
+    add_definitions(-DENABLE_WORKGRAPHS)
 endif()
 
 # reference libs used by both backends

libs/Cauldron

@@ -14,7 +14,9 @@ To build the Work Graphs Compute Rasterizer Sample, you must first install the f
 - [Visual Studio 2022](https://visualstudio.microsoft.com/downloads/)
 - [Windows 10 SDK 10.0.18362.0](https://developer.microsoft.com/en-us/windows/downloads/windows-10-sdk)
 
-Then follow the next steps:
+### Getting up and running
+
+Follow the next steps:
 
 1) Clone the repo with its submodules:
     ```
@@ -29,16 +31,12 @@ Then follow the next steps:
 
 3) Open the solution in the DX12 directory, compile and run. VK isn't supported.
 
-4) The define `ENABLE_EXPERIMENTAL_WORKGRAPHS` enables Work graphs. It's enabled by
-   default. It can be disabled if [PIX](https://devblogs.microsoft.com/pix/introduction/),
-   [RenderDoc](https://renderdoc.org) or another tool without Work graphs
-   support is used to debug. Use CMake to disable the feature:
+4) The define `ENABLE_WORKGRAPHS` enables Work graphs. It's enabled by default. It can be disabled if [PIX](https://devblogs.microsoft.com/pix/introduction/),    [RenderDoc](https://renderdoc.org) or another tool without Work graphs    support is used to debug. Use CMake to disable the feature:
     ```
     > cmake -DGFX_API_WG=OFF
     ```
 
-5) The media samples from Cauldron may change and not all scenes may load,
-   you can adjust scenes in [src\Common\GLTFSample.json](src\Common\GLTFSample.json)
+5) The media samples from Cauldron may change and not all scenes may load, you can adjust scenes in [src\Common\GLTFSample.json](src\Common\GLTFSample.json)
 
 6) You can request precise profiling data through commandline by passing json as arguments:
     ```
@@ -49,34 +47,36 @@ Then follow the next steps:
         "warmUpFrames": 50, "durationFrames": 20 } }
     ```
 
-   algorithm:\
-    0 - monolithic compute rasterizer, called once per object\
-    1 - monolithic compute rasterizer, all objects are consumed by one ExecuteIndirect\
-    2 - multi-pass compute rasterizer, all objects are consumed by a pipeline of three ExecuteIndirect\
-    3 - workgraph rasterizer, called once per object\
-    4 - workgraph rasterizer, called once in total with all object batched\
-    5 - workgraph rasterizer, all objects are consumed by an initial Broadcaster node
+   These are the custom parameters for the rasterizer, the many other options can be looked up in the Cauldron docs:
+
+   `algorithm`:
+   > 0 - monolithic compute rasterizer, called once per object
+   > 1 - monolithic compute rasterizer, all objects are consumed by one ExecuteIndirect
+   > 2 - multi-pass compute rasterizer, all objects are consumed by a pipeline of three ExecuteIndirect
+   > 3 - workgraph rasterizer, called once per object
+   > 4 - workgraph rasterizer, called once in total with all object batched
+   > 5 - workgraph rasterizer, all objects are consumed by an initial Broadcaster node
 
-   bins:\
-    1 to 15: number of buckets/bins to group triangles by, 2^(bin-1) = area
+   `bins`:
+   > 1 to 15: number of buckets/bins to group triangles by, 2^(bin-1) = area
 
-   threadlaunch:\
-    false/true: use a thread-launch or coalesce node for triangle rasterization
+   `threadlaunch`:
+   > false/true: use a thread-launch or coalesce node for triangle rasterization
 
-   profile:\
-    false/true: enables automatic profiling and generation of a csv file
+   `profile`:
+   > false/true: enables automatic profiling and generation of a csv file
 
-   filter:\
-    label: the label of the pass to collect timestamps for, if empty all passes are included in the csv
+   `filter`:
+   > label: the label of the pass to collect timestamps for, if empty all passes are included in the csv
 
-   resultsFilename:\
-    filename: the name of the csv in the current directory
+   `resultsFilename`:
+   > filename: the name of the csv in the current directory
 
-   warmUpFrames:\
-    0 to inf: how many frames to wait for initialization to settle
+   `warmUpFrames`:
+   > 0 to inf: how many frames to wait for initialization to settle
 
-   durationFrames:\
-    0 to inf: how many frames to include in the csv
+   `durationFrames`:
+   > 0 to inf: how many frames to include in the csv
 
    The values in the resulting csv are raw gpu ticks without any conversion.
    

readme.md

@@ -1,6 +1,6 @@
 // AMD SampleDX12 sample code
 // 
-// Copyright(c) 2023 Advanced Micro Devices, Inc.All rights reserved.
+// Copyright(c) 2024 Advanced Micro Devices, Inc.All rights reserved.
 // Permission is hereby granted, free of charge, to any person obtaining a copy
 // of this software and associated documentation files(the "Software"), to deal
 // in the Software without restriction, including without limitation the rights
@@ -48,8 +48,12 @@ void GLTFSample::OnParseCommandLine(LPSTR lpCmdLine, uint32_t* pWidth, uint32_t*
 
 	// ADJUSTMENT: support z-buffer rasterizer custom parameters
 	m_homogeneous = true;
+#if ENABLE_WORKGRAPHS
+	m_algorithm = 5;
+#else
 	m_algorithm = 2;
-	m_numberOfBins = 12;
+#endif
+	m_numberOfBins = 15;
 	m_fixExpansion = true;
 	m_threadLaunch = true;
 	m_smoothing = true;
@@ -500,14 +504,14 @@ void GLTFSample::OnRender()
 	Sleep(0);
 }
 
-#ifdef ENABLE_EXPERIMENTAL_WORKGRAPHS
+#ifdef ENABLE_WORKGRAPHS
 // Configure Agility SDK version
 extern "C" {
-	__declspec(dllexport) extern const UINT D3D12SDKVersion = 711;
+	__declspec(dllexport) extern const UINT D3D12SDKVersion = 613;
 }
 // Configure Agility SDK path
 extern "C" {
-	__declspec(dllexport) extern const char* D3D12SDKPath = ".\\d3d12wg\\";
+	__declspec(dllexport) extern const char* D3D12SDKPath = ".\\d3d12\\";
 }
 #endif
 
@@ -521,22 +525,7 @@ int WINAPI WinMain(HINSTANCE hInstance,
 	LPSTR lpCmdLine,
 	int nCmdShow)
 {
-	LPCSTR Name = "DX12 Work Graphs Compute Rasterizer Sample v1.0";
-
-	// Enable experimental features
-	{
-		const IID ids[] = {
-			D3D12ExperimentalShaderModels,
-#ifdef ENABLE_EXPERIMENTAL_WORKGRAPHS
-			D3D12StateObjectsExperiment
-#endif
-		};
-
-		void* configStructs[] = { nullptr, nullptr };
-		UINT configStructSizes[] = { 0, 0 };
-		ThrowIfFailed(
-			D3D12EnableExperimentalFeatures(_countof(ids), ids, configStructs, configStructSizes));
-	}
+	LPCSTR Name = "DX12 Work Graphs Compute Rasterizer Sample v1.1";
 
 	// create new DX sample
 	return RunFramework(hInstance, lpCmdLine, nCmdShow, new GLTFSample(Name));

screenshot.png

@@ -1,6 +1,6 @@
 // AMD SampleDX12 sample code
 // 
-// Copyright(c) 2023 Advanced Micro Devices, Inc.All rights reserved.
+// Copyright(c) 2024 Advanced Micro Devices, Inc.All rights reserved.
 // Permission is hereby granted, free of charge, to any person obtaining a copy
 // of this software and associated documentation files(the "Software"), to deal
 // in the Software without restriction, including without limitation the rights

src/DX12/GLTFSample.cpp

@@ -1,6 +1,6 @@
 // AMD SampleDX12 sample code
 // 
-// Copyright(c) 2023 Advanced Micro Devices, Inc.All rights reserved.
+// Copyright(c) 2024 Advanced Micro Devices, Inc.All rights reserved.
 // Permission is hereby granted, free of charge, to any person obtaining a copy
 // of this software and associated documentation files(the "Software"), to deal
 // in the Software without restriction, including without limitation the rights
@@ -31,12 +31,12 @@ void Renderer::OnCreate(Device* pDevice, SwapChain *pSwapChain, float FontSize)
 {
 	m_pDevice = pDevice;
 
-#ifdef ENABLE_EXPERIMENTAL_WORKGRAPHS
+#ifdef ENABLE_WORKGRAPHS
 	// Check work-graph support
 	//
-	D3D12_FEATURE_DATA_D3D12_OPTIONS_EXPERIMENTAL Options;
+	D3D12_FEATURE_DATA_D3D12_OPTIONS21 Options;
 	ThrowIfFailed(pDevice->GetDevice()->CheckFeatureSupport(
-		D3D12_FEATURE_D3D12_OPTIONS_EXPERIMENTAL, &Options, sizeof(Options)));
+		D3D12_FEATURE_D3D12_OPTIONS21, &Options, sizeof(Options)));
 	if (Options.WorkGraphsTier == D3D12_WORK_GRAPHS_TIER_NOT_SUPPORTED) {
 		throw std::runtime_error("Device does not report support for work graphs");
 	}
@@ -114,7 +114,7 @@ void Renderer::OnCreate(Device* pDevice, SwapChain *pSwapChain, float FontSize)
 	m_Bloom.OnCreate(pDevice, &m_ResourceViewHeaps, &m_ConstantBufferRing, &m_VidMemBufferPool, DXGI_FORMAT_R16G16B16A16_FLOAT);
 	m_TAA.OnCreate(pDevice, &m_ResourceViewHeaps, &m_VidMemBufferPool);
 	m_MagnifierPS.OnCreate(pDevice, &m_ResourceViewHeaps, &m_ConstantBufferRing, &m_VidMemBufferPool, DXGI_FORMAT_R16G16B16A16_FLOAT);
-	m_Debug.OnCreate(pDevice, &m_ResourceViewHeaps, &m_VidMemBufferPool, DXGI_FORMAT_R16G16B16A16_FLOAT);
+	m_Debug.OnCreate(pDevice, &m_ResourceViewHeaps, &m_ConstantBufferRing, &m_VidMemBufferPool, DXGI_FORMAT_R16G16B16A16_FLOAT);
 
 	// Create tonemapping pass
 	m_ToneMappingPS.OnCreate(pDevice, &m_ResourceViewHeaps, &m_ConstantBufferRing, &m_VidMemBufferPool, pSwapChain->GetFormat());
@@ -163,7 +163,7 @@ void Renderer::OnDestroy()
 	m_ResourceViewHeaps.OnDestroy();
 	m_CommandListRing.OnDestroy();
 
-#ifdef ENABLE_EXPERIMENTAL_WORKGRAPHS
+#ifdef ENABLE_WORKGRAPHS
 	m_pDevice8Native->Release();
 #endif
 }
@@ -713,7 +713,7 @@ void Renderer::OnRender(const UIState* pState, const Camera& Cam, SwapChain* pSw
 		};
 		pCmdLst1->ResourceBarrier(1, preResolve);
 
-		m_Debug.Draw(pCmdLst1, &m_VBufferSRV);
+		m_Debug.Draw(pCmdLst1, &m_VBufferSRV, pState->showBins);
 
 		D3D12_RESOURCE_BARRIER postResolve[1] = {
 			CD3DX12_RESOURCE_BARRIER::Transition(m_VBuffer.GetResource(), D3D12_RESOURCE_STATE_PIXEL_SHADER_RESOURCE, D3D12_RESOURCE_STATE_UNORDERED_ACCESS),