この節ではプッシュ定数というものを扱います。これもデスクリプタと並んで、シェーダにデータを渡す手段の1つです。
最後になりますが一番簡単なやつなので気楽に聞いてください。
プッシュ定数はデスクリプタと同じく、ユニフォーム変数のデータを渡すためのシステムです。 デスクリプタと違い、データを直接コマンドバッファに記録するのが特徴です。
後述するような制限はありますが、プッシュ定数はユニフォームバッファ+デスクリプタよりも簡単な書き方でデータを渡すことが出来ます。 処理の効率も良いらしいです。
この節では、前節まででデスクリプタを用いて渡していたデータをプッシュ定数を用いて渡すように書き換えてみます。
パイプラインレイアウトの設定
頂点入力デスクリプションやデスクリプタセットレイアウトと同じように、プッシュ定数の情報もパイプラインレイアウトに設定する必要があります。
プッシュ定数の読み込み方の情報は vk::PushConstantRange で表されます。
vk::PipelineLayoutCreateInfoの pPushConstantRangesおよびpushConstantRangeCount にvk::PushConstantRangeの配列を指定して設定します。
vk::PushConstantRange pushConstantRange[1];
pushConstantRange[0].offset = 0;
pushConstantRange[0].size = sizeof(SceneData);
pushConstantRange[0].stageFlags = vk::ShaderStageFlagBits::eVertex;
layoutCreateInfo.pPushConstantRanges = pushConstantRange;
layoutCreateInfo.pushConstantRangeCount = 1;
offsetとsizeにより、プッシュ定数として渡したバイト列からデータを取り出す位置、大きさを指定します。
stageFlagsでデータを渡す対象のシェーダを指定します。
シェーダの設定
デスクリプタによってデータを渡すときはlayoutに色々指定をしていましたが、プッシュ定数でデータを渡す場合は単にpush_constantと書くだけです。
頂点シェーダ1行目のset = 0, binding = 0の部分をpush_constantに変更してください。
layout(push_constant) uniform SceneData { // layout指定を変更
vec2 rectCenter;
} sceneData;
ここだけ変更してシェーダをコンパイルし直しておきましょう。
データの受け渡し
データを渡すためのバッファとかは何も要りません。コマンドバッファのpushConstantsメソッドに直接データを渡すだけです。
// データの用意
sceneData.rectCenter = Vec2{0.3f * cosf(time), 0.3f * sinf(time)};
time += 0.001;
// (中略)
cmdBufs[0]->pushConstants(pipelineLayout.get(), vk::ShaderStageFlagBits::eVertex, 0, sizeof(SceneData), &sceneData);
cmdBufs[0]->drawIndexed(indices.size(), 1, 0, 0, 0);
第1引数はパイプラインレイアウトです。
第2引数はデータを渡す対象とするシェーダの種類です。今回は頂点シェーダのみに渡すのでvk::ShaderStageFlagBits::eVertexを指定しています。もちろんフラグメントシェーダに渡すこともできます。
第3、第4引数は更新範囲、第5引数はプッシュ定数として用いるデータへのポインタです。
渡す処理はこれで以上です。デスクリプタがかなり煩雑だったので、それと比べると拍子抜けしてしまったのではないでしょうか。
実行してみましょう。デスクリプタのときと同じように動くはずです。
デスクリプタとの使い分け
プッシュ定数はユニフォーム変数の値をシェーダに渡す役割という点において、デスクリプタと全く変わりありません。 一体どのように使い分ければ良いのでしょう。
これは割と明確な答えがあって、小さい数値データはプッシュ定数、大きなバッファやテクスチャ画像などのデータはデスクリプタを使います。
プッシュ定数はそもそも仕様上128バイトまでしか容量を保証されていません。(参考) float型の4x4行列ならば2つ送ったら終わりです。思ったより小さいです。
注: 128バイトは最低保証であって、実際はもうちょっと大きい場合が多いと思われます。手元環境(Intel Core i7, RTX3060Ti)では256バイトでした。
実際の運用としては、カメラ座標などといったサイズが決まっている小さいデータをプッシュ定数で、モデルのテクスチャ画像やキャラクターの姿勢状態といった大きくてサイズの定まらないデータをデスクリプタで渡すことになるでしょう。
この節ではプッシュ定数の使い方について解説しました。次節、次々節では座学として、より実践的な3DCGで使われる実装について学んでいきます。この節のコード
#include <vulkan/vulkan.hpp>
#include <GLFW/glfw3.h>
#include <filesystem>
#include <fstream>
#include <iostream>
const uint32_t screenWidth = 640;
const uint32_t screenHeight = 480;
struct Vec2 {
float x, y;
};
struct Vec3 {
float x, y, z;
};
struct Vertex {
Vec2 pos;
Vec3 color;
};
std::vector<Vertex> vertices = {
Vertex{Vec2{-0.5f, -0.5f}, Vec3{0.0, 0.0, 1.0}},
Vertex{Vec2{0.5f, 0.5f}, Vec3{0.0, 1.0, 0.0}},
Vertex{Vec2{-0.5f, 0.5f}, Vec3{1.0, 0.0, 0.0}},
Vertex{Vec2{0.5f, -0.5f}, Vec3{1.0, 1.0, 1.0}},
};
std::vector<uint16_t> indices = {0, 1, 2, 1, 0, 3};
struct SceneData {
Vec2 rectCenter;
};
SceneData sceneData = {Vec2{0.3, -0.2}};
int main() {
if (!glfwInit())
return -1;
uint32_t requiredExtensionsCount;
const char **requiredExtensions = glfwGetRequiredInstanceExtensions(&requiredExtensionsCount);
vk::InstanceCreateInfo createInfo;
createInfo.enabledExtensionCount = requiredExtensionsCount;
createInfo.ppEnabledExtensionNames = requiredExtensions;
vk::UniqueInstance instance;
instance = vk::createInstanceUnique(createInfo);
glfwWindowHint(GLFW_CLIENT_API, GLFW_NO_API);
GLFWwindow *window;
window = glfwCreateWindow(screenWidth, screenHeight, "GLFW Test Window", NULL, NULL);
if (!window) {
const char *err;
glfwGetError(&err);
std::cout << err << std::endl;
glfwTerminate();
return -1;
}
VkSurfaceKHR c_surface;
auto result = glfwCreateWindowSurface(instance.get(), window, nullptr, &c_surface);
if (result != VK_SUCCESS) {
const char *err;
glfwGetError(&err);
std::cout << err << std::endl;
glfwTerminate();
return -1;
}
vk::UniqueSurfaceKHR surface{c_surface, instance.get()};
std::vector<vk::PhysicalDevice> physicalDevices = instance->enumeratePhysicalDevices();
vk::PhysicalDevice physicalDevice;
bool existsSuitablePhysicalDevice = false;
uint32_t graphicsQueueFamilyIndex;
for (size_t i = 0; i < physicalDevices.size(); i++) {
std::vector<vk::QueueFamilyProperties> queueProps = physicalDevices[i].getQueueFamilyProperties();
bool existsGraphicsQueue = false;
for (size_t j = 0; j < queueProps.size(); j++) {
if (queueProps[j].queueFlags & vk::QueueFlagBits::eGraphics && physicalDevices[i].getSurfaceSupportKHR(j, surface.get())) {
existsGraphicsQueue = true;
graphicsQueueFamilyIndex = j;
break;
}
}
std::vector<vk::ExtensionProperties> extProps = physicalDevices[i].enumerateDeviceExtensionProperties();
bool supportsSwapchainExtension = false;
for (size_t j = 0; j < extProps.size(); j++) {
if (std::string_view(extProps[j].extensionName.data()) == VK_KHR_SWAPCHAIN_EXTENSION_NAME) {
supportsSwapchainExtension = true;
break;
}
}
if (existsGraphicsQueue && supportsSwapchainExtension) {
physicalDevice = physicalDevices[i];
existsSuitablePhysicalDevice = true;
break;
}
}
if (!existsSuitablePhysicalDevice) {
std::cerr << "使用可能な物理デバイスがありません。" << std::endl;
return -1;
}
vk::DeviceCreateInfo devCreateInfo;
auto devRequiredExtensions = {VK_KHR_SWAPCHAIN_EXTENSION_NAME};
devCreateInfo.enabledExtensionCount = devRequiredExtensions.size();
devCreateInfo.ppEnabledExtensionNames = devRequiredExtensions.begin();
vk::DeviceQueueCreateInfo queueCreateInfo[1];
queueCreateInfo[0].queueFamilyIndex = graphicsQueueFamilyIndex;
queueCreateInfo[0].queueCount = 1;
float queuePriorities[1] = {1.0};
queueCreateInfo[0].pQueuePriorities = queuePriorities;
devCreateInfo.pQueueCreateInfos = queueCreateInfo;
devCreateInfo.queueCreateInfoCount = 1;
vk::UniqueDevice device = physicalDevice.createDeviceUnique(devCreateInfo);
vk::Queue graphicsQueue = device->getQueue(graphicsQueueFamilyIndex, 0);
vk::PhysicalDeviceMemoryProperties memProps = physicalDevice.getMemoryProperties();
vk::BufferCreateInfo vertBufferCreateInfo;
vertBufferCreateInfo.size = sizeof(Vertex) * vertices.size();
vertBufferCreateInfo.usage = vk::BufferUsageFlagBits::eVertexBuffer | vk::BufferUsageFlagBits::eTransferDst;
vertBufferCreateInfo.sharingMode = vk::SharingMode::eExclusive;
vk::UniqueBuffer vertexBuf = device->createBufferUnique(vertBufferCreateInfo);
vk::MemoryRequirements vertexBufMemReq = device->getBufferMemoryRequirements(vertexBuf.get());
vk::MemoryAllocateInfo vertexBufMemAllocInfo;
vertexBufMemAllocInfo.allocationSize = vertexBufMemReq.size;
bool suitableMemoryTypeFound = false;
for (uint32_t i = 0; i < memProps.memoryTypeCount; i++) {
if (vertexBufMemReq.memoryTypeBits & (1 << i) && (memProps.memoryTypes[i].propertyFlags & vk::MemoryPropertyFlagBits::eDeviceLocal)) {
vertexBufMemAllocInfo.memoryTypeIndex = i;
suitableMemoryTypeFound = true;
break;
}
}
if (!suitableMemoryTypeFound) {
std::cerr << "適切なメモリタイプが存在しません。" << std::endl;
return -1;
}
vk::UniqueDeviceMemory vertexBufMemory = device->allocateMemoryUnique(vertexBufMemAllocInfo);
device->bindBufferMemory(vertexBuf.get(), vertexBufMemory.get(), 0);
{
vk::BufferCreateInfo stagingBufferCreateInfo;
stagingBufferCreateInfo.size = sizeof(Vertex) * vertices.size();
stagingBufferCreateInfo.usage = vk::BufferUsageFlagBits::eTransferSrc;
stagingBufferCreateInfo.sharingMode = vk::SharingMode::eExclusive;
vk::UniqueBuffer stagingBuf = device->createBufferUnique(stagingBufferCreateInfo);
vk::MemoryRequirements stagingBufMemReq = device->getBufferMemoryRequirements(stagingBuf.get());
vk::MemoryAllocateInfo stagingBufMemAllocInfo;
stagingBufMemAllocInfo.allocationSize = stagingBufMemReq.size;
suitableMemoryTypeFound = false;
for (uint32_t i = 0; i < memProps.memoryTypeCount; i++) {
if (stagingBufMemReq.memoryTypeBits & (1 << i) && (memProps.memoryTypes[i].propertyFlags & vk::MemoryPropertyFlagBits::eHostVisible)) {
stagingBufMemAllocInfo.memoryTypeIndex = i;
suitableMemoryTypeFound = true;
break;
}
}
if (!suitableMemoryTypeFound) {
std::cerr << "適切なメモリタイプが存在しません。" << std::endl;
return -1;
}
vk::UniqueDeviceMemory stagingBufMemory = device->allocateMemoryUnique(stagingBufMemAllocInfo);
device->bindBufferMemory(stagingBuf.get(), stagingBufMemory.get(), 0);
void *pStagingBufMem = device->mapMemory(stagingBufMemory.get(), 0, sizeof(Vertex) * vertices.size());
std::memcpy(pStagingBufMem, vertices.data(), sizeof(Vertex) * vertices.size());
vk::MappedMemoryRange flushMemoryRange;
flushMemoryRange.memory = stagingBufMemory.get();
flushMemoryRange.offset = 0;
flushMemoryRange.size = sizeof(Vertex) * vertices.size();
device->flushMappedMemoryRanges({flushMemoryRange});
device->unmapMemory(stagingBufMemory.get());
vk::CommandPoolCreateInfo tmpCmdPoolCreateInfo;
tmpCmdPoolCreateInfo.queueFamilyIndex = graphicsQueueFamilyIndex;
tmpCmdPoolCreateInfo.flags = vk::CommandPoolCreateFlagBits::eTransient;
vk::UniqueCommandPool tmpCmdPool = device->createCommandPoolUnique(tmpCmdPoolCreateInfo);
vk::CommandBufferAllocateInfo tmpCmdBufAllocInfo;
tmpCmdBufAllocInfo.commandPool = tmpCmdPool.get();
tmpCmdBufAllocInfo.commandBufferCount = 1;
tmpCmdBufAllocInfo.level = vk::CommandBufferLevel::ePrimary;
std::vector<vk::UniqueCommandBuffer> tmpCmdBufs = device->allocateCommandBuffersUnique(tmpCmdBufAllocInfo);
vk::BufferCopy bufCopy;
bufCopy.srcOffset = 0;
bufCopy.dstOffset = 0;
bufCopy.size = sizeof(Vertex) * vertices.size();
vk::CommandBufferBeginInfo cmdBeginInfo;
cmdBeginInfo.flags = vk::CommandBufferUsageFlagBits::eOneTimeSubmit;
tmpCmdBufs[0]->begin(cmdBeginInfo);
tmpCmdBufs[0]->copyBuffer(stagingBuf.get(), vertexBuf.get(), {bufCopy});
tmpCmdBufs[0]->end();
vk::CommandBuffer submitCmdBuf[1] = {tmpCmdBufs[0].get()};
vk::SubmitInfo submitInfo;
submitInfo.commandBufferCount = 1;
submitInfo.pCommandBuffers = submitCmdBuf;
graphicsQueue.submit({submitInfo});
graphicsQueue.waitIdle();
}
vk::BufferCreateInfo indexBufferCreateInfo;
indexBufferCreateInfo.size = sizeof(uint16_t) * indices.size();
indexBufferCreateInfo.usage = vk::BufferUsageFlagBits::eIndexBuffer | vk::BufferUsageFlagBits::eTransferDst;
indexBufferCreateInfo.sharingMode = vk::SharingMode::eExclusive;
vk::UniqueBuffer indexBuf = device->createBufferUnique(indexBufferCreateInfo);
vk::MemoryRequirements indexBufMemReq = device->getBufferMemoryRequirements(indexBuf.get());
vk::MemoryAllocateInfo indexBufMemAllocInfo;
indexBufMemAllocInfo.allocationSize = indexBufMemReq.size;
suitableMemoryTypeFound = false;
for (uint32_t i = 0; i < memProps.memoryTypeCount; i++) {
if (indexBufMemReq.memoryTypeBits & (1 << i) && (memProps.memoryTypes[i].propertyFlags & vk::MemoryPropertyFlagBits::eDeviceLocal)) {
indexBufMemAllocInfo.memoryTypeIndex = i;
suitableMemoryTypeFound = true;
break;
}
}
if (!suitableMemoryTypeFound) {
std::cerr << "適切なメモリタイプが存在しません。" << std::endl;
return -1;
}
vk::UniqueDeviceMemory indexBufMemory = device->allocateMemoryUnique(indexBufMemAllocInfo);
device->bindBufferMemory(indexBuf.get(), indexBufMemory.get(), 0);
{
vk::BufferCreateInfo stagingBufferCreateInfo;
stagingBufferCreateInfo.size = sizeof(uint16_t) * indices.size();
stagingBufferCreateInfo.usage = vk::BufferUsageFlagBits::eTransferSrc;
stagingBufferCreateInfo.sharingMode = vk::SharingMode::eExclusive;
vk::UniqueBuffer stagingBuf = device->createBufferUnique(stagingBufferCreateInfo);
vk::MemoryRequirements stagingBufMemReq = device->getBufferMemoryRequirements(stagingBuf.get());
vk::MemoryAllocateInfo stagingBufMemAllocInfo;
stagingBufMemAllocInfo.allocationSize = stagingBufMemReq.size;
suitableMemoryTypeFound = false;
for (uint32_t i = 0; i < memProps.memoryTypeCount; i++) {
if (stagingBufMemReq.memoryTypeBits & (1 << i) && (memProps.memoryTypes[i].propertyFlags & vk::MemoryPropertyFlagBits::eHostVisible)) {
stagingBufMemAllocInfo.memoryTypeIndex = i;
suitableMemoryTypeFound = true;
break;
}
}
if (!suitableMemoryTypeFound) {
std::cerr << "適切なメモリタイプが存在しません。" << std::endl;
return -1;
}
vk::UniqueDeviceMemory stagingBufMemory = device->allocateMemoryUnique(stagingBufMemAllocInfo);
device->bindBufferMemory(stagingBuf.get(), stagingBufMemory.get(), 0);
void *pStagingBufMem = device->mapMemory(stagingBufMemory.get(), 0, sizeof(uint16_t) * indices.size());
std::memcpy(pStagingBufMem, indices.data(), sizeof(uint16_t) * indices.size());
vk::MappedMemoryRange flushMemoryRange;
flushMemoryRange.memory = stagingBufMemory.get();
flushMemoryRange.offset = 0;
flushMemoryRange.size = sizeof(uint16_t) * indices.size();
device->flushMappedMemoryRanges({flushMemoryRange});
device->unmapMemory(stagingBufMemory.get());
vk::CommandPoolCreateInfo tmpCmdPoolCreateInfo;
tmpCmdPoolCreateInfo.queueFamilyIndex = graphicsQueueFamilyIndex;
tmpCmdPoolCreateInfo.flags = vk::CommandPoolCreateFlagBits::eTransient;
vk::UniqueCommandPool tmpCmdPool = device->createCommandPoolUnique(tmpCmdPoolCreateInfo);
vk::CommandBufferAllocateInfo tmpCmdBufAllocInfo;
tmpCmdBufAllocInfo.commandPool = tmpCmdPool.get();
tmpCmdBufAllocInfo.commandBufferCount = 1;
tmpCmdBufAllocInfo.level = vk::CommandBufferLevel::ePrimary;
std::vector<vk::UniqueCommandBuffer> tmpCmdBufs = device->allocateCommandBuffersUnique(tmpCmdBufAllocInfo);
vk::BufferCopy bufCopy;
bufCopy.srcOffset = 0;
bufCopy.dstOffset = 0;
bufCopy.size = sizeof(uint16_t) * indices.size();
vk::CommandBufferBeginInfo cmdBeginInfo;
cmdBeginInfo.flags = vk::CommandBufferUsageFlagBits::eOneTimeSubmit;
tmpCmdBufs[0]->begin(cmdBeginInfo);
tmpCmdBufs[0]->copyBuffer(stagingBuf.get(), indexBuf.get(), {bufCopy});
tmpCmdBufs[0]->end();
vk::CommandBuffer submitCmdBuf[1] = {tmpCmdBufs[0].get()};
vk::SubmitInfo submitInfo;
submitInfo.commandBufferCount = 1;
submitInfo.pCommandBuffers = submitCmdBuf;
graphicsQueue.submit({submitInfo});
graphicsQueue.waitIdle();
}
std::vector<vk::SurfaceFormatKHR> surfaceFormats = physicalDevice.getSurfaceFormatsKHR(surface.get());
std::vector<vk::PresentModeKHR> surfacePresentModes = physicalDevice.getSurfacePresentModesKHR(surface.get());
vk::SurfaceFormatKHR swapchainFormat = surfaceFormats[0];
vk::PresentModeKHR swapchainPresentMode = surfacePresentModes[0];
vk::AttachmentDescription attachments[1];
attachments[0].format = swapchainFormat.format;
attachments[0].samples = vk::SampleCountFlagBits::e1;
attachments[0].loadOp = vk::AttachmentLoadOp::eClear;
attachments[0].storeOp = vk::AttachmentStoreOp::eStore;
attachments[0].stencilLoadOp = vk::AttachmentLoadOp::eDontCare;
attachments[0].stencilStoreOp = vk::AttachmentStoreOp::eDontCare;
attachments[0].initialLayout = vk::ImageLayout::eUndefined;
attachments[0].finalLayout = vk::ImageLayout::ePresentSrcKHR;
vk::AttachmentReference subpass0_attachmentRefs[1];
subpass0_attachmentRefs[0].attachment = 0;
subpass0_attachmentRefs[0].layout = vk::ImageLayout::eColorAttachmentOptimal;
vk::SubpassDescription subpasses[1];
subpasses[0].pipelineBindPoint = vk::PipelineBindPoint::eGraphics;
subpasses[0].colorAttachmentCount = 1;
subpasses[0].pColorAttachments = subpass0_attachmentRefs;
vk::RenderPassCreateInfo renderpassCreateInfo;
renderpassCreateInfo.attachmentCount = 1;
renderpassCreateInfo.pAttachments = attachments;
renderpassCreateInfo.subpassCount = 1;
renderpassCreateInfo.pSubpasses = subpasses;
renderpassCreateInfo.dependencyCount = 0;
renderpassCreateInfo.pDependencies = nullptr;
vk::UniqueRenderPass renderpass = device->createRenderPassUnique(renderpassCreateInfo);
vk::Viewport viewports[1];
viewports[0].x = 0.0;
viewports[0].y = 0.0;
viewports[0].minDepth = 0.0;
viewports[0].maxDepth = 1.0;
viewports[0].width = screenWidth;
viewports[0].height = screenHeight;
vk::Rect2D scissors[1];
scissors[0].offset = vk::Offset2D{0, 0};
scissors[0].extent = vk::Extent2D{screenWidth, screenHeight};
vk::PipelineViewportStateCreateInfo viewportState;
viewportState.viewportCount = 1;
viewportState.pViewports = viewports;
viewportState.scissorCount = 1;
viewportState.pScissors = scissors;
vk::VertexInputBindingDescription vertexBindingDescription[1];
vertexBindingDescription[0].binding = 0;
vertexBindingDescription[0].stride = sizeof(Vertex);
vertexBindingDescription[0].inputRate = vk::VertexInputRate::eVertex;
vk::VertexInputAttributeDescription vertexInputDescription[2];
vertexInputDescription[0].binding = 0;
vertexInputDescription[0].location = 0;
vertexInputDescription[0].format = vk::Format::eR32G32Sfloat;
vertexInputDescription[0].offset = offsetof(Vertex, pos);
vertexInputDescription[1].binding = 0;
vertexInputDescription[1].location = 1;
vertexInputDescription[1].format = vk::Format::eR32G32B32Sfloat;
vertexInputDescription[1].offset = offsetof(Vertex, color);
vk::PipelineVertexInputStateCreateInfo vertexInputInfo;
vertexInputInfo.vertexBindingDescriptionCount = std::size(vertexBindingDescription);
vertexInputInfo.pVertexBindingDescriptions = vertexBindingDescription;
vertexInputInfo.vertexAttributeDescriptionCount = std::size(vertexInputDescription);
vertexInputInfo.pVertexAttributeDescriptions = vertexInputDescription;
vk::PipelineInputAssemblyStateCreateInfo inputAssembly;
inputAssembly.topology = vk::PrimitiveTopology::eTriangleList;
inputAssembly.primitiveRestartEnable = false;
vk::PipelineRasterizationStateCreateInfo rasterizer;
rasterizer.depthClampEnable = false;
rasterizer.rasterizerDiscardEnable = false;
rasterizer.polygonMode = vk::PolygonMode::eFill;
rasterizer.lineWidth = 1.0f;
rasterizer.cullMode = vk::CullModeFlagBits::eBack;
rasterizer.frontFace = vk::FrontFace::eClockwise;
rasterizer.depthBiasEnable = false;
vk::PipelineMultisampleStateCreateInfo multisample;
multisample.sampleShadingEnable = false;
multisample.rasterizationSamples = vk::SampleCountFlagBits::e1;
vk::PipelineColorBlendAttachmentState blendattachment[1];
blendattachment[0].colorWriteMask = vk::ColorComponentFlagBits::eA | vk::ColorComponentFlagBits::eR | vk::ColorComponentFlagBits::eG | vk::ColorComponentFlagBits::eB;
blendattachment[0].blendEnable = false;
vk::PipelineColorBlendStateCreateInfo blend;
blend.logicOpEnable = false;
blend.attachmentCount = 1;
blend.pAttachments = blendattachment;
vk::PipelineLayoutCreateInfo layoutCreateInfo;
vk::PushConstantRange pushConstantRange[1];
pushConstantRange[0].offset = 0;
pushConstantRange[0].size = sizeof(SceneData);
pushConstantRange[0].stageFlags = vk::ShaderStageFlagBits::eVertex;
layoutCreateInfo.pPushConstantRanges = pushConstantRange;
layoutCreateInfo.pushConstantRangeCount = 1;
vk::UniquePipelineLayout pipelineLayout = device->createPipelineLayoutUnique(layoutCreateInfo);
size_t vertSpvFileSz = std::filesystem::file_size("shader.vert.spv");
std::ifstream vertSpvFile("shader.vert.spv", std::ios_base::binary);
std::vector<char> vertSpvFileData(vertSpvFileSz);
vertSpvFile.read(vertSpvFileData.data(), vertSpvFileSz);
vk::ShaderModuleCreateInfo vertShaderCreateInfo;
vertShaderCreateInfo.codeSize = vertSpvFileSz;
vertShaderCreateInfo.pCode = reinterpret_cast<const uint32_t *>(vertSpvFileData.data());
vk::UniqueShaderModule vertShader = device->createShaderModuleUnique(vertShaderCreateInfo);
size_t fragSpvFileSz = std::filesystem::file_size("shader.frag.spv");
std::ifstream fragSpvFile("shader.frag.spv", std::ios_base::binary);
std::vector<char> fragSpvFileData(fragSpvFileSz);
fragSpvFile.read(fragSpvFileData.data(), fragSpvFileSz);
vk::ShaderModuleCreateInfo fragShaderCreateInfo;
fragShaderCreateInfo.codeSize = fragSpvFileSz;
fragShaderCreateInfo.pCode = reinterpret_cast<const uint32_t *>(fragSpvFileData.data());
vk::UniqueShaderModule fragShader = device->createShaderModuleUnique(fragShaderCreateInfo);
vk::PipelineShaderStageCreateInfo shaderStage[2];
shaderStage[0].stage = vk::ShaderStageFlagBits::eVertex;
shaderStage[0].module = vertShader.get();
shaderStage[0].pName = "main";
shaderStage[1].stage = vk::ShaderStageFlagBits::eFragment;
shaderStage[1].module = fragShader.get();
shaderStage[1].pName = "main";
vk::GraphicsPipelineCreateInfo pipelineCreateInfo;
pipelineCreateInfo.pViewportState = &viewportState;
pipelineCreateInfo.pVertexInputState = &vertexInputInfo;
pipelineCreateInfo.pInputAssemblyState = &inputAssembly;
pipelineCreateInfo.pRasterizationState = &rasterizer;
pipelineCreateInfo.pMultisampleState = &multisample;
pipelineCreateInfo.pColorBlendState = &blend;
pipelineCreateInfo.layout = pipelineLayout.get();
pipelineCreateInfo.renderPass = renderpass.get();
pipelineCreateInfo.subpass = 0;
pipelineCreateInfo.stageCount = 2;
pipelineCreateInfo.pStages = shaderStage;
vk::UniquePipeline pipeline = device->createGraphicsPipelineUnique(nullptr, pipelineCreateInfo).value;
vk::UniqueSwapchainKHR swapchain;
std::vector<vk::Image> swapchainImages;
std::vector<vk::UniqueImageView> swapchainImageViews;
std::vector<vk::UniqueFramebuffer> swapchainFramebufs;
auto recreateSwapchain = [&]() {
swapchainFramebufs.clear();
swapchainImageViews.clear();
swapchainImages.clear();
swapchain.reset();
vk::SurfaceCapabilitiesKHR surfaceCapabilities = physicalDevice.getSurfaceCapabilitiesKHR(surface.get());
vk::SwapchainCreateInfoKHR swapchainCreateInfo;
swapchainCreateInfo.surface = surface.get();
swapchainCreateInfo.minImageCount = surfaceCapabilities.minImageCount + 1;
swapchainCreateInfo.imageFormat = swapchainFormat.format;
swapchainCreateInfo.imageColorSpace = swapchainFormat.colorSpace;
swapchainCreateInfo.imageExtent = surfaceCapabilities.currentExtent;
swapchainCreateInfo.imageArrayLayers = 1;
swapchainCreateInfo.imageUsage = vk::ImageUsageFlagBits::eColorAttachment;
swapchainCreateInfo.imageSharingMode = vk::SharingMode::eExclusive;
swapchainCreateInfo.preTransform = surfaceCapabilities.currentTransform;
swapchainCreateInfo.presentMode = swapchainPresentMode;
swapchainCreateInfo.clipped = VK_TRUE;
swapchain = device->createSwapchainKHRUnique(swapchainCreateInfo);
swapchainImages = device->getSwapchainImagesKHR(swapchain.get());
swapchainImageViews.resize(swapchainImages.size());
for (size_t i = 0; i < swapchainImages.size(); i++) {
vk::ImageViewCreateInfo imgViewCreateInfo;
imgViewCreateInfo.image = swapchainImages[i];
imgViewCreateInfo.viewType = vk::ImageViewType::e2D;
imgViewCreateInfo.format = swapchainFormat.format;
imgViewCreateInfo.components.r = vk::ComponentSwizzle::eIdentity;
imgViewCreateInfo.components.g = vk::ComponentSwizzle::eIdentity;
imgViewCreateInfo.components.b = vk::ComponentSwizzle::eIdentity;
imgViewCreateInfo.components.a = vk::ComponentSwizzle::eIdentity;
imgViewCreateInfo.subresourceRange.aspectMask = vk::ImageAspectFlagBits::eColor;
imgViewCreateInfo.subresourceRange.baseMipLevel = 0;
imgViewCreateInfo.subresourceRange.levelCount = 1;
imgViewCreateInfo.subresourceRange.baseArrayLayer = 0;
imgViewCreateInfo.subresourceRange.layerCount = 1;
swapchainImageViews[i] = device->createImageViewUnique(imgViewCreateInfo);
}
swapchainFramebufs.resize(swapchainImages.size());
for (size_t i = 0; i < swapchainImages.size(); i++) {
vk::ImageView frameBufAttachments[1];
frameBufAttachments[0] = swapchainImageViews[i].get();
vk::FramebufferCreateInfo frameBufCreateInfo;
frameBufCreateInfo.width = surfaceCapabilities.currentExtent.width;
frameBufCreateInfo.height = surfaceCapabilities.currentExtent.height;
frameBufCreateInfo.layers = 1;
frameBufCreateInfo.renderPass = renderpass.get();
frameBufCreateInfo.attachmentCount = 1;
frameBufCreateInfo.pAttachments = frameBufAttachments;
swapchainFramebufs[i] = device->createFramebufferUnique(frameBufCreateInfo);
}
};
recreateSwapchain();
vk::CommandPoolCreateInfo cmdPoolCreateInfo;
cmdPoolCreateInfo.queueFamilyIndex = graphicsQueueFamilyIndex;
cmdPoolCreateInfo.flags = vk::CommandPoolCreateFlagBits::eResetCommandBuffer;
vk::UniqueCommandPool cmdPool = device->createCommandPoolUnique(cmdPoolCreateInfo);
vk::CommandBufferAllocateInfo cmdBufAllocInfo;
cmdBufAllocInfo.commandPool = cmdPool.get();
cmdBufAllocInfo.commandBufferCount = 1;
cmdBufAllocInfo.level = vk::CommandBufferLevel::ePrimary;
std::vector<vk::UniqueCommandBuffer> cmdBufs = device->allocateCommandBuffersUnique(cmdBufAllocInfo);
vk::SemaphoreCreateInfo semaphoreCreateInfo;
vk::UniqueSemaphore swapchainImgSemaphore, imgRenderedSemaphore;
swapchainImgSemaphore = device->createSemaphoreUnique(semaphoreCreateInfo);
imgRenderedSemaphore = device->createSemaphoreUnique(semaphoreCreateInfo);
vk::FenceCreateInfo fenceCreateInfo;
fenceCreateInfo.flags = vk::FenceCreateFlagBits::eSignaled;
vk::UniqueFence imgRenderedFence = device->createFenceUnique(fenceCreateInfo);
float time = 0;
auto old = std::chrono::high_resolution_clock::now();
while (!glfwWindowShouldClose(window)) {
glfwPollEvents();
device->waitForFences({imgRenderedFence.get()}, VK_TRUE, UINT64_MAX);
vk::ResultValue acquireImgResult = device->acquireNextImageKHR(swapchain.get(), 1'000'000'000, swapchainImgSemaphore.get());
if (acquireImgResult.result == vk::Result::eSuboptimalKHR || acquireImgResult.result == vk::Result::eErrorOutOfDateKHR) {
std::cerr << "スワップチェーンを再作成します。" << std::endl;
recreateSwapchain();
continue;
}
if (acquireImgResult.result != vk::Result::eSuccess) {
std::cerr << "次フレームの取得に失敗しました。" << std::endl;
return -1;
}
device->resetFences({imgRenderedFence.get()});
sceneData.rectCenter = Vec2{0.3f * cosf(time), 0.3f * sinf(time)};
time += 0.001;
uint32_t imgIndex = acquireImgResult.value;
cmdBufs[0]->reset();
vk::CommandBufferBeginInfo cmdBeginInfo;
cmdBufs[0]->begin(cmdBeginInfo);
vk::ClearValue clearVal[1];
clearVal[0].color.float32[0] = 0.0f;
clearVal[0].color.float32[1] = 0.0f;
clearVal[0].color.float32[2] = 0.0f;
clearVal[0].color.float32[3] = 1.0f;
vk::RenderPassBeginInfo renderpassBeginInfo;
renderpassBeginInfo.renderPass = renderpass.get();
renderpassBeginInfo.framebuffer = swapchainFramebufs[imgIndex].get();
renderpassBeginInfo.renderArea = vk::Rect2D({0, 0}, {screenWidth, screenHeight});
renderpassBeginInfo.clearValueCount = 1;
renderpassBeginInfo.pClearValues = clearVal;
cmdBufs[0]->beginRenderPass(renderpassBeginInfo, vk::SubpassContents::eInline);
cmdBufs[0]->bindPipeline(vk::PipelineBindPoint::eGraphics, pipeline.get());
cmdBufs[0]->bindVertexBuffers(0, {vertexBuf.get()}, {0});
cmdBufs[0]->bindIndexBuffer(indexBuf.get(), 0, vk::IndexType::eUint16);
cmdBufs[0]->pushConstants(pipelineLayout.get(), vk::ShaderStageFlagBits::eVertex, 0, sizeof(SceneData), &sceneData);
cmdBufs[0]->drawIndexed(indices.size(), 1, 0, 0, 0);
cmdBufs[0]->endRenderPass();
cmdBufs[0]->end();
vk::CommandBuffer submitCmdBuf[1] = {cmdBufs[0].get()};
vk::SubmitInfo submitInfo;
submitInfo.commandBufferCount = 1;
submitInfo.pCommandBuffers = submitCmdBuf;
vk::Semaphore renderwaitSemaphores[] = {swapchainImgSemaphore.get()};
vk::PipelineStageFlags renderwaitStages[] = {vk::PipelineStageFlagBits::eColorAttachmentOutput};
submitInfo.waitSemaphoreCount = 1;
submitInfo.pWaitSemaphores = renderwaitSemaphores;
submitInfo.pWaitDstStageMask = renderwaitStages;
vk::Semaphore renderSignalSemaphores[] = {imgRenderedSemaphore.get()};
submitInfo.signalSemaphoreCount = 1;
submitInfo.pSignalSemaphores = renderSignalSemaphores;
graphicsQueue.submit({submitInfo}, imgRenderedFence.get());
vk::PresentInfoKHR presentInfo;
auto presentSwapchains = {swapchain.get()};
auto imgIndices = {imgIndex};
presentInfo.swapchainCount = presentSwapchains.size();
presentInfo.pSwapchains = presentSwapchains.begin();
presentInfo.pImageIndices = imgIndices.begin();
vk::Semaphore presenWaitSemaphores[] = {imgRenderedSemaphore.get()};
presentInfo.waitSemaphoreCount = 1;
presentInfo.pWaitSemaphores = presenWaitSemaphores;
graphicsQueue.presentKHR(presentInfo);
}
graphicsQueue.waitIdle();
glfwTerminate();
return 0;
}
#version 450
#extension GL_ARB_separate_shader_objects : enable
layout(push_constant) uniform SceneData {
vec2 rectCenter;
} sceneData;
layout(location = 0) in vec2 inPos;
layout(location = 1) in vec3 inColor;
layout(location = 0) out vec3 fragmentColor;
void main() {
gl_Position = vec4(sceneData.rectCenter + inPos, 0.0, 1.0);
fragmentColor = inColor;
}
#version 450
#extension GL_ARB_separate_shader_objects : enable
layout(location = 0) in vec3 fragmentColor;
layout(location = 0) out vec4 outColor;
void main() {
outColor = vec4(fragmentColor, 1.0);
}
cmake_minimum_required(VERSION 3.22)
project(vulkan-test)
set(CMAKE_CXX_STANDARD 17)
add_executable(app main.cpp)
find_package(Vulkan REQUIRED)
target_include_directories(app PRIVATE ${Vulkan_INCLUDE_DIRS})
target_link_libraries(app PRIVATE ${Vulkan_LIBRARIES})
find_package(glfw3 CONFIG REQUIRED)
target_link_libraries(app PRIVATE glfw)
公開:
最終更新: