Dynamitos/Seele
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

adding aftermath debugging and fixing forward plus

Browse Source
This commit is contained in:
Dynamitos
2021-06-04 18:27:49 +02:00
parent e00b382d4a
commit 22adb08bfc
40 changed files with 4669 additions and 112 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/Engine/Graphics/GraphicsEnums.h
+1 -1
View File
@@ -162,7 +162,7 @@ typedef uint32 KeyModifierFlags;
namespace Gfx
{
static constexpr bool useAsyncCompute = true;
static constexpr bool waitIdleOnSubmit = false;
static constexpr bool waitIdleOnSubmit = true;
static constexpr uint32 numFramesBuffered = 8;
extern uint32 currentFrameIndex;
extern double currentFrameDelta;
src/Engine/Graphics/GraphicsResources.cpp
+14 -2
View File
@@ -212,13 +212,25 @@ bool UniformBuffer::updateContents(const BulkResourceData& resourceData)
return true;
}
StructuredBuffer::StructuredBuffer(QueueFamilyMapping mapping, QueueType startQueueType)
: Buffer(mapping, startQueueType)
StructuredBuffer::StructuredBuffer(QueueFamilyMapping mapping, const BulkResourceData& resourceData)
: Buffer(mapping, resourceData.owner)
, contents(resourceData.size)
{
}
StructuredBuffer::~StructuredBuffer()
{
}
bool StructuredBuffer::updateContents(const BulkResourceData& resourceData)
{
assert(contents.size() == resourceData.size);
if(std::memcmp(contents.data(), resourceData.data, contents.size()) == 0)
{
return false;
}
std::memcpy(contents.data(), resourceData.data, contents.size());
return true;
}
VertexBuffer::VertexBuffer(QueueFamilyMapping mapping, uint32 numVertices, uint32 vertexSize, QueueType startQueueType)
: Buffer(mapping, startQueueType)
, numVertices(numVertices)
src/Engine/Graphics/GraphicsResources.h
+19 -1
View File
@@ -405,9 +405,27 @@ DEFINE_REF(IndexBuffer)
class StructuredBuffer : public Buffer
{
public:
StructuredBuffer(QueueFamilyMapping mapping, QueueType startQueueType);
StructuredBuffer(QueueFamilyMapping mapping, const BulkResourceData& bulkResourceData);
virtual ~StructuredBuffer();
virtual bool updateContents(const BulkResourceData& resourceData);
bool isDataEquals(StructuredBuffer* other)
{
if(other == nullptr)
{
return false;
}
if(contents.size() != other->contents.size())
{
return false;
}
if(std::memcmp(contents.data(), other->contents.data(), contents.size()) != 0)
{
return false;
}
return true;
}
protected:
Array<uint8> contents;
// Inherited via QueueOwnedResource
virtual void executeOwnershipBarrier(QueueType newOwner) = 0;
virtual void executePipelineBarrier(SeAccessFlags srcAccess, SePipelineStageFlags srcStage,
src/Engine/Graphics/RenderPass/BasePass.cpp
+7 -2
View File
@@ -92,7 +92,7 @@ BasePass::BasePass(PRenderGraph renderGraph, const PScene scene, Gfx::PGraphics
basePassLayout = graphics->createPipelineLayout();
lightLayout = graphics->createDescriptorLayout("LightLayout");
lightLayout->addDescriptorBinding(0, Gfx::SE_DESCRIPTOR_TYPE_STORAGE_BUFFER);
lightLayout->addDescriptorBinding(0, Gfx::SE_DESCRIPTOR_TYPE_UNIFORM_BUFFER);
lightLayout->addDescriptorBinding(1, Gfx::SE_DESCRIPTOR_TYPE_STORAGE_BUFFER);
lightLayout->addDescriptorBinding(2, Gfx::SE_DESCRIPTOR_TYPE_STORAGE_IMAGE);
lightLayout->create();
@@ -133,6 +133,10 @@ void BasePass::beginFrame()
uniformUpdate.size = sizeof(ViewParameter);
uniformUpdate.data = (uint8*)&viewParams;
viewParamBuffer->updateContents(uniformUpdate);
viewLayout->reset();
lightLayout->reset();
descriptorSets[INDEX_LIGHT_ENV] = lightLayout->allocateDescriptorSet();
descriptorSets[INDEX_VIEW_PARAMS] = viewLayout->allocateDescriptorSet();
descriptorSets[INDEX_VIEW_PARAMS]->updateBuffer(0, viewParamBuffer);
descriptorSets[INDEX_VIEW_PARAMS]->writeChanges();
for(auto &&meshBatch : scene->getStaticMeshes())
@@ -143,7 +147,6 @@ void BasePass::beginFrame()
void BasePass::render()
{
descriptorSets[INDEX_LIGHT_ENV]->updateBuffer(0, scene->getLightBuffer());
oLightIndexList->pipelineBarrier(
Gfx::SE_ACCESS_SHADER_WRITE_BIT, Gfx::SE_PIPELINE_STAGE_COMPUTE_SHADER_BIT,
@@ -151,6 +154,8 @@ void BasePass::render()
oLightGrid->pipelineBarrier(
Gfx::SE_ACCESS_SHADER_WRITE_BIT, Gfx::SE_PIPELINE_STAGE_COMPUTE_SHADER_BIT,
Gfx::SE_ACCESS_SHADER_READ_BIT, Gfx::SE_PIPELINE_STAGE_FRAGMENT_SHADER_BIT);
descriptorSets[INDEX_LIGHT_ENV]->updateBuffer(0, scene->getLightBuffer());
descriptorSets[INDEX_LIGHT_ENV]->updateBuffer(1, oLightIndexList);
descriptorSets[INDEX_LIGHT_ENV]->updateTexture(2, oLightGrid);
descriptorSets[INDEX_LIGHT_ENV]->writeChanges();
src/Engine/Graphics/RenderPass/DepthPrepass.cpp
+2 -1
View File
@@ -99,7 +99,6 @@ DepthPrepass::DepthPrepass(PRenderGraph renderGraph, const PScene scene, Gfx::PG
viewParamBuffer = graphics->createUniformBuffer(uniformInitializer);
viewLayout->create();
depthPrepassLayout->addDescriptorLayout(INDEX_VIEW_PARAMS, viewLayout);
descriptorSets[INDEX_VIEW_PARAMS] = viewLayout->allocateDescriptorSet();
primitiveLayout = graphics->createDescriptorLayout("PrimitiveLayout");
primitiveLayout->addDescriptorBinding(0, Gfx::SE_DESCRIPTOR_TYPE_UNIFORM_BUFFER);
@@ -125,6 +124,8 @@ void DepthPrepass::beginFrame()
uniformUpdate.size = sizeof(ViewParameter);
uniformUpdate.data = (uint8*)&viewParams;
viewParamBuffer->updateContents(uniformUpdate);
viewLayout->reset();
descriptorSets[INDEX_VIEW_PARAMS] = viewLayout->allocateDescriptorSet();
descriptorSets[INDEX_VIEW_PARAMS]->updateBuffer(0, viewParamBuffer);
descriptorSets[INDEX_VIEW_PARAMS]->writeChanges();
for(auto &&meshBatch : scene->getStaticMeshes())
src/Engine/Graphics/RenderPass/LightCullingPass.cpp
+33 -8
View File
@@ -23,20 +23,32 @@ LightCullingPass::~LightCullingPass()
void LightCullingPass::beginFrame()
{
uint32_t viewportWidth = viewport->getSizeX();
uint32_t viewportHeight = viewport->getSizeY();
BulkResourceData uniformUpdate;
viewParams.viewMatrix = source->getViewMatrix();
viewParams.projectionMatrix = source->getProjectionMatrix();
viewParams.cameraPosition = Vector4(source->getCameraPosition(), 0);
viewParams.inverseProjectionMatrix = glm::inverse(viewParams.projectionMatrix);
viewParams.screenDimensions = Vector2(static_cast<float>(viewport->getSizeX()), static_cast<float>(viewport->getSizeY()));
viewParams.screenDimensions = Vector2(static_cast<float>(viewportWidth), static_cast<float>(viewportHeight));
uniformUpdate.size = sizeof(ViewParameter);
uniformUpdate.data = (uint8*)&viewParams;
viewParamsBuffer->updateContents(uniformUpdate);
BulkResourceData counterReset;
uint32 reset = 0;
counterReset.data = (uint8*)&reset;
counterReset.size = sizeof(uint32);
oLightIndexCounter->updateContents(counterReset);
tLightIndexCounter->updateContents(counterReset);
cullingDescriptorLayout->reset();
lightEnvDescriptorLayout->reset();
cullingDescriptorSet = cullingDescriptorLayout->allocateDescriptorSet();
lightEnvDescriptorSet = lightEnvDescriptorLayout->allocateDescriptorSet();
cullingDescriptorSet->updateBuffer(0, viewParamsBuffer);
cullingDescriptorSet->updateBuffer(1, dispatchParamsBuffer);
cullingDescriptorSet->updateBuffer(3, frustumBuffer);
@@ -61,7 +73,6 @@ void LightCullingPass::render()
Gfx::SE_ACCESS_SHADER_READ_BIT, Gfx::SE_PIPELINE_STAGE_COMPUTE_SHADER_BIT);
depthAttachment->changeLayout(Gfx::SE_IMAGE_LAYOUT_GENERAL);
depthAttachment->transferOwnership(Gfx::QueueType::COMPUTE);
cullingDescriptorSet->updateTexture(2, depthAttachment);
cullingDescriptorSet->writeChanges();
lightEnvDescriptorSet->updateBuffer(0, scene->getLightBuffer());
@@ -85,16 +96,29 @@ void LightCullingPass::endFrame()
void LightCullingPass::publishOutputs()
{
setupFrustums();
uint32_t viewportWidth = viewport->getSizeX();
uint32_t viewportHeight = viewport->getSizeY();
glm::uvec3 numThreadGroups = glm::ceil(glm::vec3(viewportWidth / (float)BLOCK_SIZE, viewportHeight / (float)BLOCK_SIZE, 1));
dispatchParams.numThreadGroups = numThreadGroups;
dispatchParams.numThreads = numThreadGroups * glm::uvec3(BLOCK_SIZE, BLOCK_SIZE, 1);
BulkResourceData resourceData;
StructuredBufferCreateInfo createInfo;
uint32 counterReset = 0;
resourceData.size = sizeof(uint32);
resourceData.data = nullptr;
resourceData.data = (uint8*)&counterReset;
resourceData.owner = Gfx::QueueType::COMPUTE;
createInfo.bDynamic = false;
createInfo.bDynamic = true;
createInfo.resourceData = resourceData;
oLightIndexCounter = graphics->createStructuredBuffer(createInfo);
tLightIndexCounter = graphics->createStructuredBuffer(createInfo);
resourceData.size = sizeof(uint32_t) * dispatchParams.numThreadGroups.x * dispatchParams.numThreadGroups.y * dispatchParams.numThreadGroups.z * 1024;
resourceData.data = nullptr;
resourceData.size = sizeof(uint32_t)
* dispatchParams.numThreadGroups.x
* dispatchParams.numThreadGroups.y
* dispatchParams.numThreadGroups.z * 200;
createInfo.resourceData = resourceData;
createInfo.bDynamic = false;
oLightIndexList = graphics->createStructuredBuffer(createInfo);
tLightIndexList = graphics->createStructuredBuffer(createInfo);
renderGraph->registerBufferOutput("LIGHTCULLING_OLIGHTLIST", oLightIndexList);
@@ -102,7 +126,7 @@ void LightCullingPass::publishOutputs()
TextureCreateInfo textureInfo;
textureInfo.width = dispatchParams.numThreadGroups.x;
textureInfo.height = dispatchParams.numThreadGroups.y;
textureInfo.format = Gfx::SE_FORMAT_R16G16_UINT;
textureInfo.format = Gfx::SE_FORMAT_R32G32_UINT;
textureInfo.usage = Gfx::SE_IMAGE_USAGE_STORAGE_BIT;
oLightGrid = graphics->createTexture2D(textureInfo);
tLightGrid = graphics->createTexture2D(textureInfo);
@@ -139,7 +163,7 @@ void LightCullingPass::createRenderPass()
lightEnvDescriptorLayout = graphics->createDescriptorLayout("LightEnv");
//LightEnv
lightEnvDescriptorLayout->addDescriptorBinding(0, Gfx::SE_DESCRIPTOR_TYPE_STORAGE_BUFFER);
lightEnvDescriptorLayout->addDescriptorBinding(0, Gfx::SE_DESCRIPTOR_TYPE_UNIFORM_BUFFER);
cullingLayout = graphics->createPipelineLayout();
cullingLayout->addDescriptorLayout(0, cullingDescriptorLayout);
@@ -251,5 +275,6 @@ void LightCullingPass::setupFrustums()
command->dispatch(numThreadGroups.x, numThreadGroups.y, numThreadGroups.z);
Array<Gfx::PComputeCommand> commands = {command};
graphics->executeCommands(commands);
frustumBuffer->pipelineBarrier(Gfx::SE_ACCESS_SHADER_WRITE_BIT, Gfx::SE_PIPELINE_STAGE_COMPUTE_SHADER_BIT, Gfx::SE_ACCESS_SHADER_READ_BIT, Gfx::SE_PIPELINE_STAGE_COMPUTE_SHADER_BIT);
frustumBuffer->pipelineBarrier(Gfx::SE_ACCESS_SHADER_WRITE_BIT, Gfx::SE_PIPELINE_STAGE_COMPUTE_SHADER_BIT,
Gfx::SE_ACCESS_SHADER_READ_BIT, Gfx::SE_PIPELINE_STAGE_COMPUTE_SHADER_BIT);
}
src/Engine/Graphics/RenderPass/RenderPass.h
+2 -2
View File
@@ -17,13 +17,13 @@ public:
virtual void publishOutputs() = 0;
virtual void createRenderPass() = 0;
protected:
struct ViewParameter
_declspec(align(16)) struct ViewParameter
{
Matrix4 viewMatrix;
Matrix4 projectionMatrix;
Matrix4 inverseProjectionMatrix;
Vector2 screenDimensions;
Vector4 cameraPosition;
Vector2 screenDimensions;
} viewParams;
Gfx::PRenderPass renderPass;
PRenderGraph renderGraph;
src/Engine/Graphics/Vulkan/CMakeLists.txt
+5
View File
@@ -1,5 +1,10 @@
target_sources(SeeleEngine
PRIVATE
NsightAftermathGpuCrashTracker.h
NsightAftermathGpuCrashTracker.cpp
NsightAftermathHelpers.h
NsightAftermathShaderDatabase.h
NsightAftermathShaderDatabase.cpp
VulkanAllocator.h
VulkanAllocator.cpp
VulkanBuffer.cpp
src/Engine/Graphics/Vulkan/NsightAftermathGpuCrashTracker.cpp
+346
View File
@@ -0,0 +1,346 @@
//*********************************************************
//
// Copyright (c) 2019-2020, NVIDIA CORPORATION. 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 to use, copy, modify, merge, publish, distribute, sublicense,
// and/or sell copies of the Software, and to permit persons to whom the
// Software is furnished to do so, subject to the following conditions:
//
// The above copyright notice and this permission notice shall be included in
// all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
// THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
// FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
// DEALINGS IN THE SOFTWARE.
//
//*********************************************************
#include <fstream>
#include <iomanip>
#include <string>
#include "NsightAftermathGpuCrashTracker.h"
//*********************************************************
// GpuCrashTracker implementation
//*********************************************************
GpuCrashTracker::GpuCrashTracker()
: m_initialized(false)
, m_mutex()
, m_shaderDebugInfo()
, m_shaderDatabase()
{
}
GpuCrashTracker::~GpuCrashTracker()
{
// If initialized, disable GPU crash dumps
if (m_initialized)
{
GFSDK_Aftermath_DisableGpuCrashDumps();
}
}
// Initialize the GPU Crash Dump Tracker
void GpuCrashTracker::Initialize()
{
// Enable GPU crash dumps and set up the callbacks for crash dump notifications,
// shader debug information notifications, and providing additional crash
// dump description data.Only the crash dump callback is mandatory. The other two
// callbacks are optional and can be omitted, by passing nullptr, if the corresponding
// functionality is not used.
// The DeferDebugInfoCallbacks flag enables caching of shader debug information data
// in memory. If the flag is set, ShaderDebugInfoCallback will be called only
// in the event of a crash, right before GpuCrashDumpCallback. If the flag is not set,
// ShaderDebugInfoCallback will be called for every shader that is compiled.
AFTERMATH_CHECK_ERROR(GFSDK_Aftermath_EnableGpuCrashDumps(
GFSDK_Aftermath_Version_API,
GFSDK_Aftermath_GpuCrashDumpWatchedApiFlags_Vulkan,
GFSDK_Aftermath_GpuCrashDumpFeatureFlags_DeferDebugInfoCallbacks, // Let the Nsight Aftermath library cache shader debug information.
GpuCrashDumpCallback, // Register callback for GPU crash dumps.
ShaderDebugInfoCallback, // Register callback for shader debug information.
CrashDumpDescriptionCallback, // Register callback for GPU crash dump description.
this)); // Set the GpuCrashTracker object as user data for the above callbacks.
m_initialized = true;
}
// Handler for GPU crash dump callbacks from Nsight Aftermath
void GpuCrashTracker::OnCrashDump(const void* pGpuCrashDump, const uint32_t gpuCrashDumpSize)
{
// Make sure only one thread at a time...
std::lock_guard<std::mutex> lock(m_mutex);
// Write to file for later in-depth analysis with Nsight Graphics.
WriteGpuCrashDumpToFile(pGpuCrashDump, gpuCrashDumpSize);
}
// Handler for shader debug information callbacks
void GpuCrashTracker::OnShaderDebugInfo(const void* pShaderDebugInfo, const uint32_t shaderDebugInfoSize)
{
// Make sure only one thread at a time...
std::lock_guard<std::mutex> lock(m_mutex);
// Get shader debug information identifier
GFSDK_Aftermath_ShaderDebugInfoIdentifier identifier = {};
AFTERMATH_CHECK_ERROR(GFSDK_Aftermath_GetShaderDebugInfoIdentifier(
GFSDK_Aftermath_Version_API,
pShaderDebugInfo,
shaderDebugInfoSize,
&identifier));
// Store information for decoding of GPU crash dumps with shader address mapping
// from within the application.
std::vector<uint8_t> data((uint8_t*)pShaderDebugInfo, (uint8_t*)pShaderDebugInfo + shaderDebugInfoSize);
m_shaderDebugInfo[identifier].swap(data);
// Write to file for later in-depth analysis of crash dumps with Nsight Graphics
WriteShaderDebugInformationToFile(identifier, pShaderDebugInfo, shaderDebugInfoSize);
}
// Handler for GPU crash dump description callbacks
void GpuCrashTracker::OnDescription(PFN_GFSDK_Aftermath_AddGpuCrashDumpDescription addDescription)
{
// Add some basic description about the crash. This is called after the GPU crash happens, but before
// the actual GPU crash dump callback. The provided data is included in the crash dump and can be
// retrieved using GFSDK_Aftermath_GpuCrashDump_GetDescription().
addDescription(GFSDK_Aftermath_GpuCrashDumpDescriptionKey_ApplicationName, "VkHelloNsightAftermath");
addDescription(GFSDK_Aftermath_GpuCrashDumpDescriptionKey_ApplicationVersion, "v1.0");
addDescription(GFSDK_Aftermath_GpuCrashDumpDescriptionKey_UserDefined, "This is a GPU crash dump example.");
addDescription(GFSDK_Aftermath_GpuCrashDumpDescriptionKey_UserDefined + 1, "Engine State: Rendering.");
addDescription(GFSDK_Aftermath_GpuCrashDumpDescriptionKey_UserDefined + 2, "More user-defined information...");
}
// Helper for writing a GPU crash dump to a file
void GpuCrashTracker::WriteGpuCrashDumpToFile(const void* pGpuCrashDump, const uint32_t gpuCrashDumpSize)
{
// Create a GPU crash dump decoder object for the GPU crash dump.
GFSDK_Aftermath_GpuCrashDump_Decoder decoder = {};
AFTERMATH_CHECK_ERROR(GFSDK_Aftermath_GpuCrashDump_CreateDecoder(
GFSDK_Aftermath_Version_API,
pGpuCrashDump,
gpuCrashDumpSize,
&decoder));
// Use the decoder object to read basic information, like application
// name, PID, etc. from the GPU crash dump.
GFSDK_Aftermath_GpuCrashDump_BaseInfo baseInfo = {};
AFTERMATH_CHECK_ERROR(GFSDK_Aftermath_GpuCrashDump_GetBaseInfo(decoder, &baseInfo));
// Use the decoder object to query the application name that was set
// in the GPU crash dump description.
uint32_t applicationNameLength = 0;
AFTERMATH_CHECK_ERROR(GFSDK_Aftermath_GpuCrashDump_GetDescriptionSize(
decoder,
GFSDK_Aftermath_GpuCrashDumpDescriptionKey_ApplicationName,
&applicationNameLength));
std::vector<char> applicationName(applicationNameLength, '\0');
AFTERMATH_CHECK_ERROR(GFSDK_Aftermath_GpuCrashDump_GetDescription(
decoder,
GFSDK_Aftermath_GpuCrashDumpDescriptionKey_ApplicationName,
uint32_t(applicationName.size()),
applicationName.data()));
// Create a unique file name for writing the crash dump data to a file.
// Note: due to an Nsight Aftermath bug (will be fixed in an upcoming
// driver release) we may see redundant crash dumps. As a workaround,
// attach a unique count to each generated file name.
static int count = 0;
const std::string baseFileName =
std::string(applicationName.data())
+ "-"
+ std::to_string(baseInfo.pid)
+ "-"
+ std::to_string(++count);
// Write the the crash dump data to a file using the .nv-gpudmp extension
// registered with Nsight Graphics.
const std::string crashDumpFileName = baseFileName + ".nv-gpudmp";
std::ofstream dumpFile(crashDumpFileName, std::ios::out | std::ios::binary);
if (dumpFile)
{
dumpFile.write((const char*)pGpuCrashDump, gpuCrashDumpSize);
dumpFile.close();
}
// Decode the crash dump to a JSON string.
// Step 1: Generate the JSON and get the size.
uint32_t jsonSize = 0;
AFTERMATH_CHECK_ERROR(GFSDK_Aftermath_GpuCrashDump_GenerateJSON(
decoder,
GFSDK_Aftermath_GpuCrashDumpDecoderFlags_ALL_INFO,
GFSDK_Aftermath_GpuCrashDumpFormatterFlags_NONE,
ShaderDebugInfoLookupCallback,
ShaderLookupCallback,
nullptr,
ShaderSourceDebugInfoLookupCallback,
this,
&jsonSize));
// Step 2: Allocate a buffer and fetch the generated JSON.
std::vector<char> json(jsonSize);
AFTERMATH_CHECK_ERROR(GFSDK_Aftermath_GpuCrashDump_GetJSON(
decoder,
uint32_t(json.size()),
json.data()));
// Write the the crash dump data as JSON to a file.
const std::string jsonFileName = crashDumpFileName + ".json";
std::ofstream jsonFile(jsonFileName, std::ios::out | std::ios::binary);
if (jsonFile)
{
jsonFile.write(json.data(), json.size());
jsonFile.close();
}
// Destroy the GPU crash dump decoder object.
AFTERMATH_CHECK_ERROR(GFSDK_Aftermath_GpuCrashDump_DestroyDecoder(decoder));
}
// Helper for writing shader debug information to a file
void GpuCrashTracker::WriteShaderDebugInformationToFile(
GFSDK_Aftermath_ShaderDebugInfoIdentifier identifier,
const void* pShaderDebugInfo,
const uint32_t shaderDebugInfoSize)
{
// Create a unique file name.
const std::string filePath = "shader-" + std::to_string(identifier) + ".nvdbg";
std::ofstream f(filePath, std::ios::out | std::ios::binary);
if (f)
{
f.write((const char*)pShaderDebugInfo, shaderDebugInfoSize);
}
}
// Handler for shader debug information lookup callbacks.
// This is used by the JSON decoder for mapping shader instruction
// addresses to DXIL lines or HLSl source lines.
void GpuCrashTracker::OnShaderDebugInfoLookup(
const GFSDK_Aftermath_ShaderDebugInfoIdentifier& identifier,
PFN_GFSDK_Aftermath_SetData setShaderDebugInfo) const
{
// Search the list of shader debug information blobs received earlier.
auto i_debugInfo = m_shaderDebugInfo.find(identifier);
if (i_debugInfo == m_shaderDebugInfo.end())
{
// Early exit, nothing found. No need to call setShaderDebugInfo.
return;
}
// Let the GPU crash dump decoder know about the shader debug information
// that was found.
setShaderDebugInfo(i_debugInfo->second.data(), uint32_t(i_debugInfo->second.size()));
}
// Handler for shader lookup callbacks.
// This is used by the JSON decoder for mapping shader instruction
// addresses to DXIL lines or HLSL source lines.
// NOTE: If the application loads stripped shader binaries (-Qstrip_debug),
// Aftermath will require access to both the stripped and the not stripped
// shader binaries.
void GpuCrashTracker::OnShaderLookup(
const GFSDK_Aftermath_ShaderHash& shaderHash,
PFN_GFSDK_Aftermath_SetData setShaderBinary) const
{
// Find shader binary data for the shader hash in the shader database.
std::vector<uint8_t> shaderBinary;
if (!m_shaderDatabase.FindShaderBinary(shaderHash, shaderBinary))
{
// Early exit, nothing found. No need to call setShaderBinary.
return;
}
// Let the GPU crash dump decoder know about the shader data
// that was found.
setShaderBinary(shaderBinary.data(), uint32_t(shaderBinary.size()));
}
// Handler for shader source debug info lookup callbacks.
// This is used by the JSON decoder for mapping shader instruction addresses to
// HLSL source lines, if the shaders used by the application were compiled with
// separate debug info data files.
void GpuCrashTracker::OnShaderSourceDebugInfoLookup(
const GFSDK_Aftermath_ShaderDebugName& shaderDebugName,
PFN_GFSDK_Aftermath_SetData setShaderBinary) const
{
// Find source debug info for the shader DebugName in the shader database.
std::vector<uint8_t> shaderBinary;
if (!m_shaderDatabase.FindShaderBinaryWithDebugData(shaderDebugName, shaderBinary))
{
// Early exit, nothing found. No need to call setShaderBinary.
return;
}
// Let the GPU crash dump decoder know about the shader debug data that was
// found.
setShaderBinary(shaderBinary.data(), uint32_t(shaderBinary.size()));
}
// Static callback wrapper for OnCrashDump
void GpuCrashTracker::GpuCrashDumpCallback(
const void* pGpuCrashDump,
const uint32_t gpuCrashDumpSize,
void* pUserData)
{
GpuCrashTracker* pGpuCrashTracker = reinterpret_cast<GpuCrashTracker*>(pUserData);
pGpuCrashTracker->OnCrashDump(pGpuCrashDump, gpuCrashDumpSize);
}
// Static callback wrapper for OnShaderDebugInfo
void GpuCrashTracker::ShaderDebugInfoCallback(
const void* pShaderDebugInfo,
const uint32_t shaderDebugInfoSize,
void* pUserData)
{
GpuCrashTracker* pGpuCrashTracker = reinterpret_cast<GpuCrashTracker*>(pUserData);
pGpuCrashTracker->OnShaderDebugInfo(pShaderDebugInfo, shaderDebugInfoSize);
}
// Static callback wrapper for OnDescription
void GpuCrashTracker::CrashDumpDescriptionCallback(
PFN_GFSDK_Aftermath_AddGpuCrashDumpDescription addDescription,
void* pUserData)
{
GpuCrashTracker* pGpuCrashTracker = reinterpret_cast<GpuCrashTracker*>(pUserData);
pGpuCrashTracker->OnDescription(addDescription);
}
// Static callback wrapper for OnShaderDebugInfoLookup
void GpuCrashTracker::ShaderDebugInfoLookupCallback(
const GFSDK_Aftermath_ShaderDebugInfoIdentifier* pIdentifier,
PFN_GFSDK_Aftermath_SetData setShaderDebugInfo,
void* pUserData)
{
GpuCrashTracker* pGpuCrashTracker = reinterpret_cast<GpuCrashTracker*>(pUserData);
pGpuCrashTracker->OnShaderDebugInfoLookup(*pIdentifier, setShaderDebugInfo);
}
// Static callback wrapper for OnShaderLookup
void GpuCrashTracker::ShaderLookupCallback(
const GFSDK_Aftermath_ShaderHash* pShaderHash,
PFN_GFSDK_Aftermath_SetData setShaderBinary,
void* pUserData)
{
GpuCrashTracker* pGpuCrashTracker = reinterpret_cast<GpuCrashTracker*>(pUserData);
pGpuCrashTracker->OnShaderLookup(*pShaderHash, setShaderBinary);
}
// Static callback wrapper for OnShaderSourceDebugInfoLookup
void GpuCrashTracker::ShaderSourceDebugInfoLookupCallback(
const GFSDK_Aftermath_ShaderDebugName* pShaderDebugName,
PFN_GFSDK_Aftermath_SetData setShaderBinary,
void* pUserData)
{
GpuCrashTracker* pGpuCrashTracker = reinterpret_cast<GpuCrashTracker*>(pUserData);
pGpuCrashTracker->OnShaderSourceDebugInfoLookup(*pShaderDebugName, setShaderBinary);
}
src/Engine/Graphics/Vulkan/NsightAftermathGpuCrashTracker.h
+159
View File
@@ -0,0 +1,159 @@
//*********************************************************
//
// Copyright (c) 2019-2020, NVIDIA CORPORATION. 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 to use, copy, modify, merge, publish, distribute, sublicense,
// and/or sell copies of the Software, and to permit persons to whom the
// Software is furnished to do so, subject to the following conditions:
//
// The above copyright notice and this permission notice shall be included in
// all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
// THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
// FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
// DEALINGS IN THE SOFTWARE.
//
//*********************************************************
#pragma once
#include <map>
#include <mutex>
#include "NsightAftermathHelpers.h"
#include "NsightAftermathShaderDatabase.h"
//*********************************************************
// Implements GPU crash dump tracking using the Nsight
// Aftermath API.
//
class GpuCrashTracker
{
public:
GpuCrashTracker();
~GpuCrashTracker();
// Initialize the GPU crash dump tracker.
void Initialize();
private:
//*********************************************************
// Callback handlers for GPU crash dumps and related data.
//
// Handler for GPU crash dump callbacks.
void OnCrashDump(const void* pGpuCrashDump, const uint32_t gpuCrashDumpSize);
// Handler for shader debug information callbacks.
void OnShaderDebugInfo(const void* pShaderDebugInfo, const uint32_t shaderDebugInfoSize);
// Handler for GPU crash dump description callbacks.
void OnDescription(PFN_GFSDK_Aftermath_AddGpuCrashDumpDescription addDescription);
//*********************************************************
// Helpers for writing a GPU crash dump and debug information
// data to files.
//
// Helper for writing a GPU crash dump to a file.
void WriteGpuCrashDumpToFile(const void* pGpuCrashDump, const uint32_t gpuCrashDumpSize);
// Helper for writing shader debug information to a file
void WriteShaderDebugInformationToFile(
GFSDK_Aftermath_ShaderDebugInfoIdentifier identifier,
const void* pShaderDebugInfo,
const uint32_t shaderDebugInfoSize);
//*********************************************************
// Helpers for decoding GPU crash dump to JSON.
//
// Handler for shader debug info lookup callbacks.
void OnShaderDebugInfoLookup(
const GFSDK_Aftermath_ShaderDebugInfoIdentifier& identifier,
PFN_GFSDK_Aftermath_SetData setShaderDebugInfo) const;
// Handler for shader lookup callbacks.
void OnShaderLookup(
const GFSDK_Aftermath_ShaderHash& shaderHash,
PFN_GFSDK_Aftermath_SetData setShaderBinary) const;
// Handler for shader instructions lookup callbacks.
void OnShaderInstructionsLookup(
const GFSDK_Aftermath_ShaderInstructionsHash& shaderInstructionsHash,
PFN_GFSDK_Aftermath_SetData setShaderBinary) const;
// Handler for shader source debug info lookup callbacks.
void OnShaderSourceDebugInfoLookup(
const GFSDK_Aftermath_ShaderDebugName& shaderDebugName,
PFN_GFSDK_Aftermath_SetData setShaderBinary) const;
//*********************************************************
// Static callback wrappers.
//
// GPU crash dump callback.
static void GpuCrashDumpCallback(
const void* pGpuCrashDump,
const uint32_t gpuCrashDumpSize,
void* pUserData);
// Shader debug information callback.
static void ShaderDebugInfoCallback(
const void* pShaderDebugInfo,
const uint32_t shaderDebugInfoSize,
void* pUserData);
// GPU crash dump description callback.
static void CrashDumpDescriptionCallback(
PFN_GFSDK_Aftermath_AddGpuCrashDumpDescription addDescription,
void* pUserData);
// Shader debug information lookup callback.
static void ShaderDebugInfoLookupCallback(
const GFSDK_Aftermath_ShaderDebugInfoIdentifier* pIdentifier,
PFN_GFSDK_Aftermath_SetData setShaderDebugInfo,
void* pUserData);
// Shader lookup callback.
static void ShaderLookupCallback(
const GFSDK_Aftermath_ShaderHash* pShaderHash,
PFN_GFSDK_Aftermath_SetData setShaderBinary,
void* pUserData);
// Shader instructions lookup callback.
static void ShaderInstructionsLookupCallback(
const GFSDK_Aftermath_ShaderInstructionsHash* pShaderInstructionsHash,
PFN_GFSDK_Aftermath_SetData setShaderBinary,
void* pUserData);
// Shader source debug info lookup callback.
static void ShaderSourceDebugInfoLookupCallback(
const GFSDK_Aftermath_ShaderDebugName* pShaderDebugName,
PFN_GFSDK_Aftermath_SetData setShaderBinary,
void* pUserData);
//*********************************************************
// GPU crash tracker state.
//
// Is the GPU crash dump tracker initialized?
bool m_initialized;
// For thread-safe access of GPU crash tracker state.
mutable std::mutex m_mutex;
// List of Shader Debug Information by ShaderDebugInfoIdentifier.
std::map<GFSDK_Aftermath_ShaderDebugInfoIdentifier, std::vector<uint8_t>> m_shaderDebugInfo;
// The mock shader database.
ShaderDatabase m_shaderDatabase;
};
src/Engine/Graphics/Vulkan/NsightAftermathHelpers.h
+142
View File
@@ -0,0 +1,142 @@
//*********************************************************
//
// Copyright (c) 2019, NVIDIA CORPORATION. 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 to use, copy, modify, merge, publish, distribute, sublicense,
// and/or sell copies of the Software, and to permit persons to whom the
// Software is furnished to do so, subject to the following conditions:
//
// The above copyright notice and this permission notice shall be included in
// all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
// THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
// FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
// DEALINGS IN THE SOFTWARE.
//
//*********************************************************
#pragma once
#include <iomanip>
#include <string>
#include <sstream>
#include <vulkan/vulkan.hpp>
#include "GFSDK_Aftermath.h"
#include "GFSDK_Aftermath_GpuCrashDump.h"
#include "GFSDK_Aftermath_GpuCrashDumpDecoding.h"
//*********************************************************
// Some std::to_string overloads for some Nsight Aftermath
// API types.
//
namespace std
{
template<typename T>
inline std::string to_hex_string(T n)
{
std::stringstream stream;
stream << std::setfill('0') << std::setw(2 * sizeof(T)) << std::hex << n;
return stream.str();
}
inline std::string to_string(GFSDK_Aftermath_Result result)
{
return std::string("0x") + to_hex_string(static_cast<uint32_t>(result));
}
inline std::string to_string(const GFSDK_Aftermath_ShaderDebugInfoIdentifier& identifier)
{
return to_hex_string(identifier.id[0]) + "-" + to_hex_string(identifier.id[1]);
}
inline std::string to_string(const GFSDK_Aftermath_ShaderHash& hash)
{
return to_hex_string(hash.hash);
}
inline std::string to_string(const GFSDK_Aftermath_ShaderInstructionsHash& hash)
{
return to_hex_string(hash.hash) + "-" + to_hex_string(hash.hash);
}
} // namespace std
//*********************************************************
// Helper for comparing shader hashes and debug info identifier.
//
// Helper for comparing GFSDK_Aftermath_ShaderDebugInfoIdentifier.
inline bool operator<(const GFSDK_Aftermath_ShaderDebugInfoIdentifier& lhs, const GFSDK_Aftermath_ShaderDebugInfoIdentifier& rhs)
{
if (lhs.id[0] == rhs.id[0])
{
return lhs.id[1] < rhs.id[1];
}
return lhs.id[0] < rhs.id[0];
}
// Helper for comparing GFSDK_Aftermath_ShaderHash.
inline bool operator<(const GFSDK_Aftermath_ShaderHash& lhs, const GFSDK_Aftermath_ShaderHash& rhs)
{
return lhs.hash < rhs.hash;
}
// Helper for comparing GFSDK_Aftermath_ShaderInstructionsHash.
inline bool operator<(const GFSDK_Aftermath_ShaderInstructionsHash& lhs, const GFSDK_Aftermath_ShaderInstructionsHash& rhs)
{
return lhs.hash < rhs.hash;
}
// Helper for comparing GFSDK_Aftermath_ShaderDebugName.
inline bool operator<(const GFSDK_Aftermath_ShaderDebugName& lhs, const GFSDK_Aftermath_ShaderDebugName& rhs)
{
return strncmp(lhs.name, rhs.name, sizeof(lhs.name)) < 0;
}
//*********************************************************
// Helper for checking Nsight Aftermath failures.
//
inline std::string AftermathErrorMessage(GFSDK_Aftermath_Result result)
{
switch (result)
{
case GFSDK_Aftermath_Result_FAIL_DriverVersionNotSupported:
return "Unsupported driver version - requires a recent NVIDIA R445 display driver or newer.";
default:
return "Aftermath Error 0x" + std::to_hex_string(result);
}
}
// Helper macro for checking Nsight Aftermath results and throwing exception
// in case of a failure.
#ifdef _WIN32
#define AFTERMATH_CHECK_ERROR(FC) \
[&]() { \
GFSDK_Aftermath_Result _result = FC; \
if (!GFSDK_Aftermath_SUCCEED(_result)) \
{ \
std::cout << AftermathErrorMessage(_result).c_str() << std::endl; \
exit(1); \
} \
}()
#else
#define AFTERMATH_CHECK_ERROR(FC) \
[&]() { \
GFSDK_Aftermath_Result _result = FC; \
if (!GFSDK_Aftermath_SUCCEED(_result)) \
{ \
printf("%s\n", AftermathErrorMessage(_result).c_str()); \
fflush(stdout); \
exit(1); \
} \
}()
#endif
src/Engine/Graphics/Vulkan/NsightAftermathShaderDatabase.cpp
+147
View File
@@ -0,0 +1,147 @@
//*********************************************************
//
// Copyright (c) 2019-2020, NVIDIA CORPORATION. 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 to use, copy, modify, merge, publish, distribute, sublicense,
// and/or sell copies of the Software, and to permit persons to whom the
// Software is furnished to do so, subject to the following conditions:
//
// The above copyright notice and this permission notice shall be included in
// all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
// THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
// FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
// DEALINGS IN THE SOFTWARE.
//
//*********************************************************
#include <fstream>
#include <iomanip>
#include "NsightAftermathShaderDatabase.h"
//*********************************************************
// ShaderDatabase implementation
//*********************************************************
ShaderDatabase::ShaderDatabase()
: m_shaderBinaries()
, m_shaderBinariesWithDebugInfo()
{
// Add shader binaries to database
AddShaderBinary("cube.vert.spirv");
AddShaderBinary("cube.frag.spirv");
// Add the not stripped shader binaries to the database, too.
AddShaderBinaryWithDebugInfo("cube.vert.spirv", "cube.vert.full.spirv");
AddShaderBinaryWithDebugInfo("cube.frag.spirv", "cube.frag.full.spirv");
}
ShaderDatabase::~ShaderDatabase()
{
}
bool ShaderDatabase::ReadFile(const char* filename, std::vector<uint8_t>& data)
{
std::ifstream fs(filename, std::ios::in | std::ios::binary);
if (!fs)
{
return false;
}
fs.seekg(0, std::ios::end);
data.resize(fs.tellg());
fs.seekg(0, std::ios::beg);
fs.read(reinterpret_cast<char*>(data.data()), data.size());
fs.close();
return true;
}
void ShaderDatabase::AddShaderBinary(const char* shaderFilePath)
{
// Read the shader binary code from the file
std::vector<uint8_t> data;
if (!ReadFile(shaderFilePath, data))
{
return;
}
// Create shader hash for the shader
const GFSDK_Aftermath_SpirvCode shader{ data.data(), uint32_t(data.size()) };
GFSDK_Aftermath_ShaderHash shaderHash;
AFTERMATH_CHECK_ERROR(GFSDK_Aftermath_GetShaderHashSpirv(
GFSDK_Aftermath_Version_API,
&shader,
&shaderHash));
// Store the data for shader mapping when decoding GPU crash dumps.
// cf. FindShaderBinary()
m_shaderBinaries[shaderHash].swap(data);
}
void ShaderDatabase::AddShaderBinaryWithDebugInfo(const char* strippedShaderFilePath, const char* shaderFilePath)
{
// Read the shader debug data from the file
std::vector<uint8_t> data;
if (!ReadFile(shaderFilePath, data))
{
return;
}
std::vector<uint8_t> strippedData;
if (!ReadFile(strippedShaderFilePath, strippedData))
{
return;
}
// Generate shader debug name.
GFSDK_Aftermath_ShaderDebugName debugName;
const GFSDK_Aftermath_SpirvCode shader{ data.data(), uint32_t(data.size()) };
const GFSDK_Aftermath_SpirvCode strippedShader{ strippedData.data(), uint32_t(strippedData.size()) };
AFTERMATH_CHECK_ERROR(GFSDK_Aftermath_GetShaderDebugNameSpirv(
GFSDK_Aftermath_Version_API,
&shader,
&strippedShader,
&debugName));
// Store the data for shader instruction address mapping when decoding GPU crash dumps.
// cf. FindShaderBinaryWithDebugData()
m_shaderBinariesWithDebugInfo[debugName].swap(data);
}
// Find a shader binary by shader hash.
bool ShaderDatabase::FindShaderBinary(const GFSDK_Aftermath_ShaderHash& shaderHash, std::vector<uint8_t>& shader) const
{
// Find shader binary data for the shader hash
auto i_shader = m_shaderBinaries.find(shaderHash);
if (i_shader == m_shaderBinaries.end())
{
// Nothing found.
return false;
}
shader = i_shader->second;
return true;
}
// Find a shader binary with debug information by shader debug name.
bool ShaderDatabase::FindShaderBinaryWithDebugData(const GFSDK_Aftermath_ShaderDebugName& shaderDebugName, std::vector<uint8_t>& shader) const
{
// Find shader binary for the shader debug name.
auto i_shader = m_shaderBinariesWithDebugInfo.find(shaderDebugName);
if (i_shader == m_shaderBinariesWithDebugInfo.end())
{
// Nothing found.
return false;
}
shader = i_shader->second;
return true;
}
src/Engine/Graphics/Vulkan/NsightAftermathShaderDatabase.h
+64
View File
@@ -0,0 +1,64 @@
//*********************************************************
//
// Copyright (c) 2019, NVIDIA CORPORATION. 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 to use, copy, modify, merge, publish, distribute, sublicense,
// and/or sell copies of the Software, and to permit persons to whom the
// Software is furnished to do so, subject to the following conditions:
//
// The above copyright notice and this permission notice shall be included in
// all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
// THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
// FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
// DEALINGS IN THE SOFTWARE.
//
//*********************************************************
#pragma once
#include <vector>
#include <map>
#include <mutex>
#include "NsightAftermathHelpers.h"
//*********************************************************
// Implements a very simple shader database to help demonstrate
// how to use the Nsight Aftermath GPU crash dump decoder API.
//
// In a real world scenario this would be part of an offline
// analysis tool. This is for demonstration purposes only!
//
class ShaderDatabase
{
public:
ShaderDatabase();
~ShaderDatabase();
// Find a shader bytecode binary by shader hash.
bool FindShaderBinary(const GFSDK_Aftermath_ShaderHash& shaderHash, std::vector<uint8_t>& shader) const;
// Find a source shader debug info by shader debug name generated by the DXC compiler.
bool FindShaderBinaryWithDebugData(const GFSDK_Aftermath_ShaderDebugName& shaderDebugName, std::vector<uint8_t>& shader) const;
private:
void AddShaderBinary(const char* shaderFilePath);
void AddShaderBinaryWithDebugInfo(const char* strippedShaderFilePath, const char* shaderFilePath);
static bool ReadFile(const char* filename, std::vector<uint8_t>& data);
// List of shader binaries by ShaderHash.
std::map<GFSDK_Aftermath_ShaderHash, std::vector<uint8_t>> m_shaderBinaries;
// List of available shader binaries with source debug information by ShaderDebugName.
std::map<GFSDK_Aftermath_ShaderDebugName, std::vector<uint8_t>> m_shaderBinariesWithDebugInfo;
};
src/Engine/Graphics/Vulkan/VulkanBuffer.cpp
+39 -1
View File
@@ -350,7 +350,7 @@ VkAccessFlags UniformBuffer::getDestAccessMask()
}
StructuredBuffer::StructuredBuffer(PGraphics graphics, const StructuredBufferCreateInfo &resourceData)
: Gfx::StructuredBuffer(graphics->getFamilyMapping(), resourceData.resourceData.owner)
: Gfx::StructuredBuffer(graphics->getFamilyMapping(), resourceData.resourceData)
, Vulkan::ShaderBuffer(graphics, resourceData.resourceData.size, VK_BUFFER_USAGE_STORAGE_BUFFER_BIT, currentOwner, resourceData.bDynamic)
{
if (resourceData.resourceData.data != nullptr)
@@ -365,6 +365,44 @@ StructuredBuffer::~StructuredBuffer()
{
}
bool StructuredBuffer::updateContents(const BulkResourceData &resourceData)
{
Gfx::StructuredBuffer::updateContents(resourceData);
//We always want to update, as the contents could be different on the GPU
void* data = lock();
std::memcpy(data, resourceData.data, resourceData.size);
unlock();
return true;
}
void* StructuredBuffer::lock(bool bWriteOnly)
{
if(dedicatedStagingBuffer != nullptr)
{
return dedicatedStagingBuffer->getMappedPointer();
}
return ShaderBuffer::lock(bWriteOnly);
}
void StructuredBuffer::unlock()
{
if(dedicatedStagingBuffer != nullptr)
{
dedicatedStagingBuffer->flushMappedMemory();
PCmdBuffer cmdBuffer = graphics->getQueueCommands(currentOwner)->getCommands();
VkCommandBuffer cmdHandle = cmdBuffer->getHandle();
VkBufferCopy region;
std::memset(&region, 0, sizeof(VkBufferCopy));
region.size = ShaderBuffer::size;
vkCmdCopyBuffer(cmdHandle, dedicatedStagingBuffer->getHandle(), buffers[currentBuffer].buffer, 1, &region);
graphics->getQueueCommands(currentOwner)->submitCommands();
}
else
{
ShaderBuffer::unlock();
}
}
void StructuredBuffer::requestOwnershipTransfer(Gfx::QueueType newOwner)
{
Gfx::QueueOwnedResource::transferOwnership(newOwner);
src/Engine/Graphics/Vulkan/VulkanCommandBuffer.cpp
+20 -14
View File
@@ -91,13 +91,12 @@ void CmdBuffer::executeCommands(const Array<Gfx::PRenderCommand>& commands)
for (uint32 i = 0; i < commands.size(); ++i)
{
auto command = commands[i].cast<RenderCommand>();
// Cache array and size to save on pointer access
for(auto boundDescriptor : command->boundDescriptors)
{
boundDescriptor->currentlyBound = this;
}
command->end();
executingRenders.add(command);
for(auto descriptor : command->boundDescriptors)
{
boundDescriptors.add(descriptor);
}
cmdBuffers[i] = command->getHandle();
}
vkCmdExecuteCommands(handle, (uint32)cmdBuffers.size(), cmdBuffers.data());
@@ -109,13 +108,12 @@ void CmdBuffer::executeCommands(const Array<Gfx::PComputeCommand>& commands)
for (uint32 i = 0; i < commands.size(); ++i)
{
auto command = commands[i].cast<ComputeCommand>();
// Cache array and size to save on pointer access
for(auto boundDescriptor : command->boundDescriptors)
{
boundDescriptor->currentlyBound = this;
}
command->end();
executingComputes.add(command);
for(auto descriptor : command->boundDescriptors)
{
boundDescriptors.add(descriptor);
}
cmdBuffers[i] = command->getHandle();
}
vkCmdExecuteCommands(handle, (uint32)cmdBuffers.size(), cmdBuffers.data());
@@ -145,6 +143,10 @@ void CmdBuffer::refreshFence()
command->reset();
}
executingRenders.clear();
for(auto descriptor : boundDescriptors)
{
descriptor->unbind();
}
state = State::ReadyBegin;
}
}
@@ -196,10 +198,6 @@ void SecondaryCmdBuffer::end()
void SecondaryCmdBuffer::reset()
{
vkResetCommandBuffer(handle, VK_COMMAND_BUFFER_RESET_RELEASE_RESOURCES_BIT);
for(auto boundDescriptor : boundDescriptors)
{
boundDescriptor->currentlyBound = nullptr;
}
boundDescriptors.clear();
ready = true;
}
@@ -251,6 +249,8 @@ void RenderCommand::bindDescriptor(Gfx::PDescriptorSet descriptorSet)
{
auto descriptor = descriptorSet.cast<DescriptorSet>();
boundDescriptors.add(descriptor.getHandle());
descriptor->bind();
//std::cout << "Binding descriptor " << descriptor->getHandle() << " to cmd " << handle << std::endl;
VkDescriptorSet setHandle = descriptor->getHandle();
vkCmdBindDescriptorSets(handle, VK_PIPELINE_BIND_POINT_GRAPHICS, pipeline->getLayout(), descriptorSet->getSetIndex(), 1, &setHandle, 0, nullptr);
}
@@ -260,6 +260,8 @@ void RenderCommand::bindDescriptor(const Array<Gfx::PDescriptorSet>& descriptorS
for(uint32 i = 0; i < descriptorSets.size(); ++i)
{
auto descriptorSet = descriptorSets[i].cast<DescriptorSet>();
descriptorSet->bind();
//std::cout << "Binding descriptor " << descriptorSet->getHandle() << " to cmd " << handle << std::endl;
boundDescriptors.add(descriptorSet.getHandle());
sets[descriptorSet->getSetIndex()] = descriptorSet->getHandle();
}
@@ -326,6 +328,8 @@ void ComputeCommand::bindDescriptor(Gfx::PDescriptorSet descriptorSet)
{
auto descriptor = descriptorSet.cast<DescriptorSet>();
boundDescriptors.add(descriptor.getHandle());
descriptor->bind();
//std::cout << "Binding descriptor " << descriptor->getHandle() << " to cmd " << handle << std::endl;
VkDescriptorSet setHandle = descriptor->getHandle();
vkCmdBindDescriptorSets(handle, VK_PIPELINE_BIND_POINT_COMPUTE, pipeline->getLayout(), descriptorSet->getSetIndex(), 1, &setHandle, 0, nullptr);
}
@@ -337,6 +341,8 @@ void ComputeCommand::bindDescriptor(const Array<Gfx::PDescriptorSet>& descriptor
{
auto descriptorSet = descriptorSets[i].cast<DescriptorSet>();
boundDescriptors.add(descriptorSet.getHandle());
descriptorSet->bind();
//std::cout << "Binding descriptor " << descriptorSet->getHandle() << " to cmd " << handle << std::endl;
sets[descriptorSet->getSetIndex()] = descriptorSet->getHandle();
}
vkCmdBindDescriptorSets(handle, VK_PIPELINE_BIND_POINT_COMPUTE, pipeline->getLayout(), 0, (uint32)descriptorSets.size(), sets, 0, nullptr);
src/Engine/Graphics/Vulkan/VulkanCommandBuffer.h
+2 -1
View File
@@ -31,6 +31,7 @@ DEFINE_REF(CmdBufferBase)
DECLARE_REF(RenderCommand)
DECLARE_REF(ComputeCommand)
DECLARE_REF(DescriptorSet)
DECLARE_REF(CommandBufferManager)
class CmdBuffer : public CmdBufferBase
{
@@ -68,6 +69,7 @@ private:
Array<VkPipelineStageFlags> waitFlags;
Array<PRenderCommand> executingRenders;
Array<PComputeCommand> executingComputes;
Array<DescriptorSet*> boundDescriptors;
friend class RenderCommand;
friend class CommandBufferManager;
friend class Queue;
@@ -76,7 +78,6 @@ DEFINE_REF(CmdBuffer)
DECLARE_REF(GraphicsPipeline)
DECLARE_REF(ComputePipeline)
DECLARE_REF(DescriptorSet)
class SecondaryCmdBuffer: public CmdBufferBase
{
src/Engine/Graphics/Vulkan/VulkanDescriptorSets.cpp
+1 -2
View File
@@ -220,7 +220,7 @@ void DescriptorSet::writeChanges()
{
if(isCurrentlyBound())
{
currentlyBound->waitForCommand(1000000000u);
std::cout << "Descriptor currently bound, allocate a new one instead" << std::endl;
assert(!isCurrentlyBound());
}
vkUpdateDescriptorSets(graphics->getDevice(), (uint32)writeDescriptors.size(), writeDescriptors.data(), 0, nullptr);
@@ -289,7 +289,6 @@ void DescriptorAllocator::allocateDescriptorSet(Gfx::PDescriptorSet &descriptorS
{
//If it hasnt been initialized, allocate it
VK_CHECK(vkAllocateDescriptorSets(graphics->getDevice(), &allocInfo, &cachedHandles[setIndex]->setHandle));
}
cachedHandles[setIndex]->currentlyInUse = true;
descriptorSet = cachedHandles[setIndex];
src/Engine/Graphics/Vulkan/VulkanDescriptorSets.h
+12 -3
View File
@@ -64,7 +64,7 @@ public:
, graphics(graphics)
, owner(owner)
, currentlyInUse(false)
, currentlyBound(nullptr)
, currentlyBound(false)
{
}
virtual ~DescriptorSet();
@@ -77,12 +77,20 @@ public:
inline bool isCurrentlyBound() const
{
return currentlyBound != nullptr;
return currentlyBound;
}
inline bool isCurrentlyInUse() const
{
return currentlyInUse;
}
void bind()
{
currentlyBound = true;
}
void unbind()
{
currentlyBound = false;
}
void free()
{
currentlyInUse = false;
@@ -104,7 +112,8 @@ private:
VkDescriptorSet setHandle;
PGraphics graphics;
PDescriptorAllocator owner;
PCmdBuffer currentlyBound;
//PCmdBuffer currentlyBound;
bool currentlyBound;
bool currentlyInUse;
friend class DescriptorAllocator;
friend class CmdBuffer;
src/Engine/Graphics/Vulkan/VulkanGraphics.cpp
+12
View File
@@ -412,6 +412,8 @@ void Graphics::setupDebugCallback()
VK_DEBUG_REPORT_ERROR_BIT_EXT | VK_DEBUG_REPORT_WARNING_BIT_EXT | VK_DEBUG_REPORT_PERFORMANCE_WARNING_BIT_EXT);
VK_CHECK(CreateDebugReportCallbackEXT(instance, &createInfo, nullptr, &callback));
crashTracker.Initialize();
}
void Graphics::pickPhysicalDevice()
@@ -537,6 +539,16 @@ void Graphics::createDevice(GraphicsInitializer initializer)
queueInfos.data(),
(uint32)queueInfos.size(),
&features);
#if ENABLE_VALIDATION
VkDeviceDiagnosticsConfigCreateInfoNV crashDiagInfo;
crashDiagInfo.sType = VK_STRUCTURE_TYPE_DEVICE_DIAGNOSTICS_CONFIG_CREATE_INFO_NV;
crashDiagInfo.pNext = nullptr;
crashDiagInfo.flags =
VK_DEVICE_DIAGNOSTICS_CONFIG_ENABLE_RESOURCE_TRACKING_BIT_NV |
VK_DEVICE_DIAGNOSTICS_CONFIG_ENABLE_AUTOMATIC_CHECKPOINTS_BIT_NV |
VK_DEVICE_DIAGNOSTICS_CONFIG_ENABLE_SHADER_DEBUG_INFO_BIT_NV;
deviceInfo.pNext = &crashDiagInfo;
#endif
deviceInfo.enabledExtensionCount = (uint32)initializer.deviceExtensions.size();
deviceInfo.ppEnabledExtensionNames = initializer.deviceExtensions.data();
deviceInfo.enabledLayerCount = (uint32_t)initializer.layers.size();
src/Engine/Graphics/Vulkan/VulkanGraphics.h
+2
View File
@@ -1,6 +1,7 @@
#pragma once
#include "VulkanGraphicsResources.h"
#include "Graphics/Graphics.h"
#include "NsightAftermathGpuCrashTracker.h"
namespace Seele
{
@@ -98,6 +99,7 @@ protected:
Map<uint32, PFramebuffer> allocatedFramebuffers;
PAllocator allocator;
PStagingManager stagingManager;
GpuCrashTracker crashTracker;
friend class Window;
};
src/Engine/Graphics/Vulkan/VulkanGraphicsEnums.h
+4
View File
@@ -8,6 +8,10 @@
VkResult res = (f); \
if (res != VK_SUCCESS) \
{ \
if(res == VK_ERROR_DEVICE_LOST) \
{ \
std::this_thread::sleep_for(std::chrono::seconds(3)); \
} \
std::cout << "Fatal : VkResult is \"" << res << "\" in " << __FILE__ << " at line " << __LINE__ << std::endl; \
assert(res == VK_SUCCESS); \
} \
src/Engine/Graphics/Vulkan/VulkanGraphicsResources.h
+3
View File
@@ -160,7 +160,10 @@ class StructuredBuffer : public Gfx::StructuredBuffer, public ShaderBuffer
public:
StructuredBuffer(PGraphics graphics, const StructuredBufferCreateInfo &resourceData);
virtual ~StructuredBuffer();
virtual bool updateContents(const BulkResourceData &resourceData);
virtual void* lock(bool bWriteOnly = true) override;
virtual void unlock() override;
protected:
// Inherited via Vulkan::Buffer
virtual VkAccessFlags getSourceAccessMask();
src/Engine/Graphics/Vulkan/VulkanPipelineCache.cpp
+4
View File
@@ -370,7 +370,11 @@ PComputePipeline PipelineCache::createPipeline(const ComputePipelineCreateInfo&
computeStage->getModuleHandle(),
computeStage->getEntryPointName());
VkPipeline pipelineHandle;
auto beginTime = std::chrono::high_resolution_clock::now();
VK_CHECK(vkCreateComputePipelines(graphics->getDevice(), cache, 1, &createInfo, nullptr, &pipelineHandle));
auto endTime = std::chrono::high_resolution_clock::now();
int64 delta = std::chrono::duration_cast<std::chrono::microseconds>(endTime - beginTime).count();
std::cout << "Compute creation time: " << delta << std::endl;
PComputePipeline result = new ComputePipeline(graphics, pipelineHandle, layout, computeInfo);
return result;
}
src/Engine/Graphics/Vulkan/VulkanQueue.cpp
+36 -36
View File
@@ -22,44 +22,44 @@ Queue::~Queue()
void Queue::submitCommandBuffer(PCmdBuffer cmdBuffer, uint32 numSignalSemaphores, VkSemaphore *signalSemaphores)
{
std::scoped_lock lck(queueLock);
assert(cmdBuffer->state == CmdBuffer::State::Ended);
PFence fence = cmdBuffer->fence;
assert(!fence->isSignaled());
const VkCommandBuffer cmdBuffers[] = {cmdBuffer->handle};
VkSubmitInfo submitInfo =
init::SubmitInfo();
submitInfo.commandBufferCount = 1;
submitInfo.pCommandBuffers = cmdBuffers;
submitInfo.signalSemaphoreCount = static_cast<uint32>(numSignalSemaphores);
submitInfo.pSignalSemaphores = signalSemaphores;
Array<VkSemaphore> waitSemaphores;
if (cmdBuffer->waitSemaphores.size() > 0)
{
std::scoped_lock lck(queueLock);
assert(cmdBuffer->state == CmdBuffer::State::Ended);
PFence fence = cmdBuffer->fence;
assert(!fence->isSignaled());
const VkCommandBuffer cmdBuffers[] = {cmdBuffer->handle};
VkSubmitInfo submitInfo =
init::SubmitInfo();
submitInfo.commandBufferCount = 1;
submitInfo.pCommandBuffers = cmdBuffers;
submitInfo.signalSemaphoreCount = static_cast<uint32>(numSignalSemaphores);
submitInfo.pSignalSemaphores = signalSemaphores;
Array<VkSemaphore> waitSemaphores;
if (cmdBuffer->waitSemaphores.size() > 0)
for (PSemaphore semaphore : cmdBuffer->waitSemaphores)
{
for (PSemaphore semaphore : cmdBuffer->waitSemaphores)
{
waitSemaphores.add(semaphore->getHandle());
}
submitInfo.waitSemaphoreCount = static_cast<uint32>(cmdBuffer->waitSemaphores.size());
submitInfo.pWaitSemaphores = waitSemaphores.data();
submitInfo.pWaitDstStageMask = cmdBuffer->waitFlags.data();
waitSemaphores.add(semaphore->getHandle());
}
VK_CHECK(vkQueueSubmit(queue, 1, &submitInfo, fence->getHandle()));
cmdBuffer->state = CmdBuffer::State::Submitted;
cmdBuffer->waitFlags.clear();
cmdBuffer->waitSemaphores.clear();
if (Gfx::waitIdleOnSubmit)
{
fence->wait(200 * 1000ull);
}
cmdBuffer->refreshFence();
graphics->getStagingManager()->clearPending();
submitInfo.waitSemaphoreCount = static_cast<uint32>(cmdBuffer->waitSemaphores.size());
submitInfo.pWaitSemaphores = waitSemaphores.data();
submitInfo.pWaitDstStageMask = cmdBuffer->waitFlags.data();
}
VK_CHECK(vkQueueSubmit(queue, 1, &submitInfo, fence->getHandle()));
cmdBuffer->state = CmdBuffer::State::Submitted;
cmdBuffer->waitFlags.clear();
cmdBuffer->waitSemaphores.clear();
if (Gfx::waitIdleOnSubmit)
{
fence->wait(200 * 1000ull);
}
cmdBuffer->refreshFence();
graphics->getStagingManager()->clearPending();
}