device/compute/compute.h - GCC Code Coverage Report

device/compute/compute.h

Line	Branch	Exec	Source
1			// SPDX-License-Identifier: BSD-3-Clause
2			#pragma once
3
4			#include <algorithm>
5			#include <array>
6			#include <concepts>
7			#include <cstdint>
8			#include <deque>
9			#include <memory>
10			#include <mutex>
11			#include <optional>
12			#include <span>
13			#include <string>
14			#include <string_view>
15			#include <utility>
16			#include <variant>
17			#include <vector>
18
19			#include <fmt/format.h>
20
21			#include "chrono_batch_render/device/buffer.h"
22			#include "chrono_batch_render/device/commands.h"
23			#include "chrono_batch_render/device/compute/runtime_compiler.h"
24			#include "chrono_batch_render/device/device.h"
25			#include "chrono_batch_render/device/queue.h"
26			#include "chrono_batch_render/device/slang/compiler.h"
27			#include "chrono_batch_render/device/texture.h"
28			#include "chrono_batch_render/util/disk_cache.h"
29			#include "chrono_batch_render/util/error.h"
30			#include "chrono_batch_render/util/profiler.h"
31
32			namespace chrono::batch_render::device::compute {
33
34			constexpr uint32_t kMaxParamBytes = 4096;
35			constexpr unsigned kMaxAnisotropy = 16;
36
37		✗	struct ComputeDevice {
38			Backend backend;
39			uint32_t ordinal;
40			std::string name;
41			std::array<uint8_t, 16> uuid;
42			uint64_t memory;
43			bool integrated;
44			uint32_t warpSize;
45			uint32_t maxThreads;
46			bool textures;
47			};
48
49		✗	inline AdapterInfo DescribeCompute(const ComputeDevice& device) {
50		✗	AdapterInfo adapter;
51		✗	adapter.backend = device.backend;
52		✗	adapter.ordinal = device.ordinal;
53	0/1 ✗ Branch 1 not taken.	✗	adapter.name = device.name;
54		✗	adapter.uuid = device.uuid;
55		✗	adapter.memoryBytes = device.memory;
56	0/2 ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	adapter.type = device.integrated ? AdapterType::Integrated : AdapterType::Discrete;
57
58		✗	Features& features = adapter.features;
59		✗	features.unifiedMemory = device.integrated;
60		✗	features.hostMappedMemory = device.integrated;
61		✗	features.externalMemory = true;
62		✗	features.kernelStatistics = true;
63		✗	features.hardwareTextures = device.textures;
64		✗	features.anisotropicFiltering = device.textures;
65		✗	features.float64 = true;
66		✗	features.atomicFloat32Add = true;
67		✗	features.atomicInt64 = true;
68		✗	features.shaderClock = true;
69		✗	features.waveSizeMin = features.waveSizeMax = device.warpSize;
70
71		✗	adapter.limits.maxBufferBytes = device.memory;
72		✗	adapter.limits.maxGroupSize = device.maxThreads;
73		✗	adapter.limits.maxParamBytes = kMaxParamBytes;
74	0/2 ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	adapter.limits.maxAnisotropy = device.textures ? float(kMaxAnisotropy) : 1.0f;
75		✗	adapter.limits.maxReservedBytes = adapter.limits.maxBufferBytes;
76		✗	return adapter;
77		✗	}
78
79			constexpr uint64_t kPoolKeptBytes = 256ull * 1024 * 1024;
80
81			template <class T>
82			concept ComputeTraits =
83			requires(const T::Device& device, const T::Api& api, T::Stream stream, T::Event event,
84			T::Graph graph, T::Pool pool, uint64_t address, void* host, const void* source,
85			uint64_t bytes, uint32_t value, bool flag, T::Module module, T::Function function,
86			const std::string& binary, Size3 groups, const T::ArrayDesc& array,
87			const T::ResourceDesc& resource, const T::SamplerDesc& sampler) {
88			{ device.Functions() } -> std::same_as<const typename T::Api&>;
89			T::Enter(device);
90
91			{ T::Allocate(api, bytes) } -> std::same_as<uint64_t>;
92			{ T::Free(api, address) } -> std::same_as<void>;
93			{ T::CreatePool(device, bytes) } -> std::same_as<typename T::Pool>;
94			{ T::DestroyPool(api, pool) } -> std::same_as<void>;
95			{ T::AllocateFrom(api, pool, bytes, stream) } -> std::same_as<uint64_t>;
96			{ T::FreeTo(api, address, stream) } -> std::same_as<void>;
97			{ T::SynchronizeStream(api, stream) } -> std::same_as<void>;
98			{ T::AllocateHost(api, bytes, flag) } -> std::same_as<void*>;
99			{ T::DeviceAddress(api, host) } -> std::same_as<uint64_t>;
100			{ T::FreeHost(api, host) } -> std::same_as<void>;
101
102			{ T::CreateStream(api) } -> std::same_as<typename T::Stream>;
103			{ T::DestroyStream(api, stream) } -> std::same_as<void>;
104			{ T::CreateEvent(api, flag) } -> std::same_as<typename T::Event>;
105			{ T::DestroyEvent(api, event) } -> std::same_as<void>;
106			{ T::Record(api, event, stream, flag) } -> std::same_as<void>;
107			{ T::WaitEvent(api, stream, event) } -> std::same_as<void>;
108			{ T::ApplicationStream(host) } -> std::same_as<typename T::Stream>;
109			{ T::ElapsedNs(api, event, event) } -> std::same_as<uint64_t>;
110			{ T::Query(api, event) } -> std::same_as<bool>;
111			{ T::Synchronize(api, event) } -> std::same_as<void>;
112			{ T::Copy(api, address, address, bytes, stream) } -> std::same_as<void>;
113			{ T::Fill(api, address, value, bytes, stream) } -> std::same_as<void>;
114			{ T::Upload(api, address, source, bytes, stream) } -> std::same_as<void>;
115			{ T::CreateTextureArray(api, array, value) } -> std::same_as<void*>;
116			{ T::DestroyTextureArray(api, host) } -> std::same_as<void>;
117			{ T::TextureLevel(api, host, value) } -> std::same_as<void*>;
118			{ T::CreateTextureObject(api, resource, sampler) } -> std::same_as<uint64_t>;
119			{ T::DestroyTextureObject(api, address) } -> std::same_as<void>;
120			{ T::CopyToArray(api, host, address, bytes, value, stream) } -> std::same_as<void>;
121			T::BeginCapture(api, stream);
122			{ T::EndCapture(api, stream) } -> std::same_as<typename T::Graph>;
123			{ T::LaunchGraph(api, graph, stream) } -> std::same_as<void>;
124			{ T::DestroyGraph(api, graph) } -> std::same_as<void>;
125
126			{ T::RuntimeCompiler() } -> std::same_as<const RuntimeCompiler&>;
127			{ T::Prelude() } -> std::same_as<std::string_view>;
128			{
129			T::CompileOptions(binary, groups, value, flag)
130			} -> std::same_as<std::vector<std::string>>;
131			{ T::LoadModule(api, binary) } -> std::same_as<typename T::Module>;
132			{ T::UnloadModule(api, module) } -> std::same_as<void>;
133			{ T::Symbol(api, module, binary.c_str()) } -> std::same_as<std::pair<uint64_t, size_t>>;
134			{ T::CopyToDevice(api, address, source, bytes) } -> std::same_as<void>;
135			{ T::GetFunction(api, module, binary.c_str()) } -> std::same_as<typename T::Function>;
136			{ T::Stats(api, function) } -> std::same_as<KernelStats>;
137			{ T::Launch(api, function, groups, groups, &host, stream) } -> std::same_as<void>;
138			{ T::kParamBuffer } -> std::convertible_to<uintptr_t>;
139			{ T::kParamBufferSize } -> std::convertible_to<uintptr_t>;
140			{ T::kParamEnd } -> std::convertible_to<uintptr_t>;
141			};
142
143			template <ComputeTraits Traits>
144		✗	void ExternalCopy(const typename Traits::Device& device, BufferRange dst, uint64_t source,
145			Stream producer) {
146		✗	const auto scope = Traits::Enter(device);
147	0/4 void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::cuda::Traits>(chrono::batch_render::device::cuda::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 0 not taken. ✗ Branch 1 not taken. void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::hip::Traits>(chrono::batch_render::device::hip::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	const typename Traits::Api& api = device.Functions();
148	0/4 void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::cuda::Traits>(chrono::batch_render::device::cuda::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 0 not taken. ✗ Branch 1 not taken. void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::hip::Traits>(chrono::batch_render::device::hip::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (producer.IsCompute()) {
149	0/2 void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::cuda::Traits>(chrono::batch_render::device::cuda::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 1 not taken. void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::hip::Traits>(chrono::batch_render::device::hip::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 1 not taken.	✗	Traits::Copy(api, dst.Address(), source, dst.Bytes(),
150			Traits::ApplicationStream(producer.Handle()));
151			return;
152			}
153	0/4 void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::cuda::Traits>(chrono::batch_render::device::cuda::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 1 not taken. ✗ Branch 4 not taken. void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::hip::Traits>(chrono::batch_render::device::hip::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 1 not taken. ✗ Branch 4 not taken.	✗	const typename Traits::Stream stream = Traits::CreateStream(api);
154	0/2 void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::cuda::Traits>(chrono::batch_render::device::cuda::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 1 not taken. void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::hip::Traits>(chrono::batch_render::device::hip::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 1 not taken.	✗	Traits::Copy(api, dst.Address(), source, dst.Bytes(), stream);
155	0/2 void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::cuda::Traits>(chrono::batch_render::device::cuda::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 1 not taken. void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::hip::Traits>(chrono::batch_render::device::hip::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 1 not taken.	✗	Traits::SynchronizeStream(api, stream);
156	0/2 void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::cuda::Traits>(chrono::batch_render::device::cuda::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 1 not taken. void chrono::batch_render::device::compute::ExternalCopy<chrono::batch_render::device::hip::Traits>(chrono::batch_render::device::hip::Traits::Device const&, chrono::batch_render::device::BufferRange, unsigned long, chrono::batch_render::Stream): ✗ Branch 1 not taken.	✗	Traits::DestroyStream(api, stream);
157		✗	}
158
159			template <ComputeTraits Traits> class ComputeQueue;
160
161			template <ComputeTraits Traits> class ComputeBuffer : public Buffer {
162			public:
163			using Device = Traits::Device;
164
165		✗	static Ref<Buffer> Create(Device& device, const BufferDesc& desc) {
166		✗	const auto scope = Traits::Enter(device);
167	0/6 chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::cuda::Traits>::Create(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken. ✗ Branch 3 not taken. ✗ Branch 4 not taken. chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::hip::Traits>::Create(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken. ✗ Branch 3 not taken. ✗ Branch 4 not taken.	✗	Ref<ComputeBuffer> buffer(new ComputeBuffer(device, desc));
168		✗	const auto& api = buffer->Functions();
169	0/4 chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::cuda::Traits>::Create(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::hip::Traits>::Create(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	switch (desc.memory) {
170		✗	case Memory::Device:
171	0/2 chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::cuda::Traits>::Create(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::hip::Traits>::Create(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken.	✗	buffer->m_queue = &static_cast<ComputeQueue<Traits>&>(device.GetQueue());
172	0/2 chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::cuda::Traits>::Create(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::hip::Traits>::Create(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken.	✗	buffer->m_address = buffer->m_queue->Allocate(desc.bytes);
173		✗	break;
174		✗	default:
175	0/4 chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::cuda::Traits>::Create(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken. ✗ Branch 4 not taken. chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::hip::Traits>::Create(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken. ✗ Branch 4 not taken.	✗	buffer->m_mapped = Traits::AllocateHost(api, desc.bytes, desc.memory == Memory::Upload);
176	0/2 chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::cuda::Traits>::Create(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::hip::Traits>::Create(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken.	✗	buffer->m_address = Traits::DeviceAddress(api, buffer->m_mapped);
177		✗	break;
178			}
179	0/4 chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::cuda::Traits>::Create(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::hip::Traits>::Create(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	return buffer;
180		✗	}
181
182		✗	~ComputeBuffer() override {
183		✗	const auto scope = Traits::Enter(Owner());
184	0/4 chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::cuda::Traits>::~ComputeBuffer(): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::hip::Traits>::~ComputeBuffer(): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (m_mapped != nullptr) {
185		✗	Traits::FreeHost(Functions(), m_mapped);
186	0/4 chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::cuda::Traits>::~ComputeBuffer(): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::hip::Traits>::~ComputeBuffer(): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	} else if (m_address != 0) {
187		✗	m_queue->Free(m_address);
188			}
189		✗	}
190
191			private:
192	0/2 chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::cuda::Traits>::Create(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::hip::Traits>::Create(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken.	✗	ComputeBuffer(Device& device, const BufferDesc& desc) : Buffer(device, desc) {}
193
194			const ComputeQueue<Traits>* m_queue = nullptr;
195
196		✗	const Device& Owner() const {
197		✗	return static_cast<const Device&>(GetDevice());
198			}
199
200	0/4 chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::cuda::Traits>::Create(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken. ✗ Branch 2 not taken. chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::hip::Traits>::Create(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken. ✗ Branch 2 not taken.	✗	const Traits::Api& Functions() const {
201	0/4 chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::cuda::Traits>::Create(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken. ✗ Branch 2 not taken. chrono::batch_render::device::compute::ComputeBuffer<chrono::batch_render::device::hip::Traits>::Create(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::BufferDesc const&): ✗ Branch 1 not taken. ✗ Branch 2 not taken.	✗	return Owner().Functions();
202			}
203			};
204
205			struct ArrayFormat {
206			int format;
207			unsigned channels;
208			};
209
210		✗	inline ArrayFormat TextureArrayFormat(const TextureDesc& desc) {
211		✗	constexpr int kUnsigned8 = 0x01;
212		✗	constexpr int kUnsigned16 = 0x02;
213		✗	constexpr int kHalf = 0x10;
214		✗	constexpr int kFloat = 0x20;
215		✗	constexpr ArrayFormat kFormats[] = {{kUnsigned8, 1}, {kUnsigned8, 2}, {kUnsigned8, 4},
216			{kUnsigned16, 1}, {kUnsigned16, 2}, {kUnsigned16, 4},
217			{kHalf, 1}, {kHalf, 2}, {kHalf, 4},
218			{kFloat, 1}, {kFloat, 2}, {kFloat, 4}};
219	0/2 ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (TexelBlockSide(desc.format) != 1) {
220		✗	throw UnsupportedError(
221	0/1 ✗ Branch 2 not taken.	✗	"texture \"{}\" is block compressed, which texture objects can't hold", desc.label);
222			}
223		✗	return kFormats[uint32_t(desc.format)];
224			}
225
226			template <ComputeTraits Traits> class ComputeTexture : public Texture {
227			public:
228			using Device = Traits::Device;
229
230		✗	ComputeTexture(Device& device, const TextureDesc& desc)
231	0/4 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::ComputeTexture(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::TextureDesc const&): ✗ Branch 2 not taken. ✗ Branch 5 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::ComputeTexture(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::TextureDesc const&): ✗ Branch 2 not taken. ✗ Branch 5 not taken.	✗	: Texture(device, desc), m_format(TextureArrayFormat(desc)) {
232	0/2 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::ComputeTexture(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::TextureDesc const&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::ComputeTexture(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::TextureDesc const&): ✗ Branch 1 not taken.	✗	const auto scope = Traits::Enter(device);
233	0/2 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::ComputeTexture(chrono::batch_render::device::cuda::CudaDevice&, chrono::batch_render::device::TextureDesc const&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::ComputeTexture(chrono::batch_render::device::hip::HipDevice&, chrono::batch_render::device::TextureDesc const&): ✗ Branch 1 not taken.	✗	m_records = Traits::Allocate(Functions(), sizeof(m_objects));
234			try {
235		✗	m_array = Traits::CreateTextureArray(Functions(),
236		✗	{.Width = desc.width,
237		✗	.Height = desc.height,
238		✗	.Format = m_format.format,
239		✗	.NumChannels = m_format.channels},
240		✗	desc.levels);
241		✗	} catch (...) {
242		✗	Traits::Free(Functions(), m_records);
243		✗	throw;
244			}
245		✗	}
246
247		✗	~ComputeTexture() override {
248		✗	const auto scope = Traits::Enter(Owner());
249	0/4 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::~ComputeTexture(): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::~ComputeTexture(): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	for (const TextureObjects& objects : m_objects) {
250	0/4 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::~ComputeTexture(): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::~ComputeTexture(): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (objects.filtered != 0) {
251		✗	Traits::DestroyTextureObject(Functions(), objects.filtered);
252		✗	Traits::DestroyTextureObject(Functions(), objects.texels);
253			}
254			}
255		✗	Traits::Free(Functions(), m_records);
256		✗	Traits::DestroyTextureArray(Functions(), m_array);
257		✗	}
258
259		✗	uint64_t KernelHandle(TextureWrap wrapS, TextureWrap wrapT) override {
260		✗	const uint32_t index = TextureHandle(0, wrapS, wrapT);
261		✗	const uint64_t address = m_records + uint64_t(index) * sizeof(TextureObjects);
262		✗	const std::lock_guard lock(m_mutex);
263	0/4 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	TextureObjects& objects = m_objects[index];
264	0/4 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (objects.filtered == 0) {
265	0/2 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 1 not taken.	✗	const auto scope = Traits::Enter(Owner());
266	0/2 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 1 not taken.	✗	const TextureDesc& desc = Desc();
267		✗	constexpr int kMipmappedArray = 1;
268		✗	typename Traits::ResourceDesc resource{.resType = kMipmappedArray};
269		✗	resource.res.mipmap.hMipmappedArray = m_array;
270	0/2 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 1 not taken.	✗	objects = {.filtered = Traits::CreateTextureObject(Functions(), resource,
271		✗	Sampler(wrapS, wrapT, true)),
272	0/2 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 1 not taken.	✗	.texels = Traits::CreateTextureObject(Functions(), resource,
273		✗	Sampler(wrapS, wrapT, false)),
274		✗	.width = desc.width,
275		✗	.height = desc.height,
276		✗	.numLevels = desc.levels,
277		✗	.channels = m_format.channels};
278	0/2 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::KernelHandle(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap): ✗ Branch 1 not taken.	✗	Traits::CopyToDevice(Functions(), address, &objects, sizeof(objects));
279		✗	}
280		✗	return address;
281		✗	}
282
283		✗	void Copy(uint32_t level, uint64_t source, Traits::Stream stream) const {
284	0/4 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::Copy(unsigned int, unsigned long, chrono::batch_render::device::cuda::CUstream_st) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::Copy(unsigned int, unsigned long, chrono::batch_render::device::hip::ihipStream_t) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	const TextureDesc& desc = Desc();
285		✗	Traits::CopyToArray(Functions(), Traits::TextureLevel(Functions(), m_array, level), source,
286	0/4 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::Copy(unsigned int, unsigned long, chrono::batch_render::device::cuda::CUstream_st) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::Copy(unsigned int, unsigned long, chrono::batch_render::device::hip::ihipStream_t) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	uint64_t(TextureLevelSide(desc.width, level)) *
287		✗	TexelBlockBytes(desc.format),
288	0/4 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::Copy(unsigned int, unsigned long, chrono::batch_render::device::cuda::CUstream_st) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::Copy(unsigned int, unsigned long, chrono::batch_render::device::hip::ihipStream_t) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	TextureLevelSide(desc.height, level), stream);
289		✗	}
290
291			private:
292			ArrayFormat m_format;
293			void* m_array = nullptr;
294			uint64_t m_records = 0;
295			std::array<TextureObjects, kTextureSamplers> m_objects{};
296			std::mutex m_mutex;
297
298		✗	const Device& Owner() const {
299	0/2 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::Copy(unsigned int, unsigned long, chrono::batch_render::device::cuda::CUstream_st) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::Copy(unsigned int, unsigned long, chrono::batch_render::device::hip::ihipStream_t) const: ✗ Branch 1 not taken.	✗	return static_cast<const Device&>(GetDevice());
300			}
301
302	0/12 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::Copy(unsigned int, unsigned long, chrono::batch_render::device::cuda::CUstream_st) const: ✗ Branch 3 not taken. ✗ Branch 6 not taken. ✗ Branch 9 not taken. ✗ Branch 16 not taken. ✗ Branch 19 not taken. ✗ Branch 22 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::Copy(unsigned int, unsigned long, chrono::batch_render::device::hip::ihipStream_t) const: ✗ Branch 3 not taken. ✗ Branch 6 not taken. ✗ Branch 9 not taken. ✗ Branch 16 not taken. ✗ Branch 19 not taken. ✗ Branch 22 not taken.	✗	const Traits::Api& Functions() const {
303	0/12 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::Copy(unsigned int, unsigned long, chrono::batch_render::device::cuda::CUstream_st) const: ✗ Branch 3 not taken. ✗ Branch 6 not taken. ✗ Branch 9 not taken. ✗ Branch 16 not taken. ✗ Branch 19 not taken. ✗ Branch 22 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::Copy(unsigned int, unsigned long, chrono::batch_render::device::hip::ihipStream_t) const: ✗ Branch 3 not taken. ✗ Branch 6 not taken. ✗ Branch 9 not taken. ✗ Branch 16 not taken. ✗ Branch 19 not taken. ✗ Branch 22 not taken.	✗	return Owner().Functions();
304			}
305
306		✗	Traits::SamplerDesc Sampler(TextureWrap wrapS, TextureWrap wrapT, bool filtered) const {
307		✗	constexpr int kPoint = 0;
308		✗	constexpr int kLinear = 1;
309		✗	constexpr unsigned kNormalizedCoordinates = 0x02;
310		✗	constexpr unsigned kSrgb = 0x10;
311		✗	const auto address = [](TextureWrap wrap) {
312		✗	constexpr int kModes[] = {0, 2, 1, 3};
313		✗	return kModes[uint32_t(wrap)];
314			};
315	0/4 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::Sampler(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap, bool) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::Sampler(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap, bool) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	const TextureDesc& desc = Desc();
316		✗	return {.addressMode = {address(wrapS), address(wrapT), address(wrapT)},
317		✗	.filterMode = filtered ? kLinear : kPoint,
318		✗	.flags = kNormalizedCoordinates \|
319		✗	(desc.colorSpace == TexelColorSpace::Srgb ? kSrgb : 0u),
320		✗	.maxAnisotropy = filtered ? kMaxAnisotropy : 1,
321			.mipmapFilterMode = filtered ? kLinear : kPoint,
322	0/12 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::Sampler(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap, bool) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. ✗ Branch 2 not taken. ✗ Branch 3 not taken. ✗ Branch 4 not taken. ✗ Branch 5 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::Sampler(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap, bool) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. ✗ Branch 2 not taken. ✗ Branch 3 not taken. ✗ Branch 4 not taken. ✗ Branch 5 not taken.	✗	.maxMipmapLevelClamp = float(desc.levels - 1)};
323			}
324			};
325
326		✗	template <ComputeTraits Traits> struct LoadedModule {
327			Traits::Module module{};
328			uint64_t dispatchArgs = 0;
329			uint64_t counts = 0;
330			};
331
332			template <ComputeTraits Traits> class ComputeKernel : public Kernel {
333			public:
334			using Device = Traits::Device;
335
336	0/2 chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::cuda::Traits>::Sampler(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap, bool) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeTexture<chrono::batch_render::device::hip::Traits>::Sampler(chrono::batch_render::TextureWrap, chrono::batch_render::TextureWrap, bool) const: ✗ Branch 1 not taken.	✗	explicit ComputeKernel(Device& owner) : m_owner(owner) {}
337
338		✗	void Launch(const Traits::Api& api, Size3 groups, uint64_t counts, const void* params,
339			uint32_t bytes, Traits::Stream stream) const {
340	0/4 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Launch(chrono::batch_render::device::cuda::Api const&, chrono::batch_render::device::Size3, unsigned long, void const, unsigned int, chrono::batch_render::device::cuda::CUstream_st) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Launch(chrono::batch_render::device::hip::Api const&, chrono::batch_render::device::Size3, unsigned long, void const, unsigned int, chrono::batch_render::device::hip::ihipStream_t) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	uint64_t& written = m_module->counts;
341	0/4 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Launch(chrono::batch_render::device::cuda::Api const&, chrono::batch_render::device::Size3, unsigned long, void const, unsigned int, chrono::batch_render::device::cuda::CUstream_st) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Launch(chrono::batch_render::device::hip::Api const&, chrono::batch_render::device::Size3, unsigned long, void const, unsigned int, chrono::batch_render::device::hip::ihipStream_t) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (counts != written) {
342		✗	Traits::Fill(api, m_module->dispatchArgs, uint32_t(counts), 4, stream);
343		✗	Traits::Fill(api, m_module->dispatchArgs + 4, uint32_t(counts >> 32), 4, stream);
344		✗	written = counts;
345			}
346		✗	size_t size = bytes;
347		✗	void* extra[] = {reinterpret_cast<void>(Traits::kParamBuffer), const_cast<void>(params),
348			reinterpret_cast<void*>(Traits::kParamBufferSize), &size,
349			reinterpret_cast<void*>(Traits::kParamEnd)};
350	0/4 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Launch(chrono::batch_render::device::cuda::Api const&, chrono::batch_render::device::Size3, unsigned long, void const, unsigned int, chrono::batch_render::device::cuda::CUstream_st) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Launch(chrono::batch_render::device::hip::Api const&, chrono::batch_render::device::Size3, unsigned long, void const, unsigned int, chrono::batch_render::device::hip::ihipStream_t) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	Traits::Launch(api, m_function, groups, GroupSize(), bytes > 0 ? extra : nullptr, stream);
351		✗	}
352
353		✗	uint64_t& Counts() const {
354		✗	return m_module->counts;
355			}
356
357			static constexpr uint64_t kUnknownCounts = ~uint64_t(0);
358
359			protected:
360		✗	KernelStats Load(const KernelCode& code, const DiskCache& cache,
361			const DeviceGlobals& globals) override {
362		✗	const RuntimeCompiler& compiler = Traits::RuntimeCompiler();
363		✗	const std::string& architecture = m_owner.Adapter().architecture;
364		✗	const std::vector<std::string> options = Traits::CompileOptions(
365		✗	architecture, code.shape.groupSize, code.shape.maxRegisters, code.debugInfo);
366		✗	const SourceHeader prelude{kPreludeHeader, Traits::Prelude()};
367	0/4 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&): ✗ Branch 1 not taken. ✗ Branch 4 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&): ✗ Branch 1 not taken. ✗ Branch 4 not taken.	✗	const RuntimeCompiler::Compiled binary =
368			compiler.CompileCached(code.code, {&prelude, 1}, options,
369	0/2 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&): ✗ Branch 1 not taken.	✗	fmt::format("kernel {} of program \"{}\" for {}", Name(),
370	0/2 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&): ✗ Branch 1 not taken.	✗	GetProgram().Label(), architecture),
371			cache);
372
373	0/2 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&): ✗ Branch 1 not taken.	✗	const auto scope = Traits::Enter(m_owner);
374	0/2 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&): ✗ Branch 1 not taken.	✗	const auto& api = m_owner.Functions();
375	0/10 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&): ✗ Branch 1 not taken. ✗ Branch 4 not taken. ✗ Branch 5 not taken. ✗ Branch 6 not taken. ✗ Branch 7 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&): ✗ Branch 1 not taken. ✗ Branch 4 not taken. ✗ Branch 5 not taken. ✗ Branch 6 not taken. ✗ Branch 7 not taken.	✗	m_module =
376	0/2 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const: ✗ Branch 1 not taken.	✗	std::static_pointer_cast<LoadedModule<Traits>>(m_owner.Modules().Find(binary.code, [&] {
377		✗	Device& owner = m_owner;
378		✗	const std::shared_ptr<LoadedModule<Traits>> loaded(
379		✗	new LoadedModule<Traits>{}, [&owner](LoadedModule<Traits>* unloading) {
380	0/4 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (unloading->module != Module{}) {
381		✗	const auto unloadScope = Traits::Enter(owner);
382	0/2 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 1 not taken.	✗	Traits::UnloadModule(owner.Functions(), unloading->module);
383		✗	}
384	0/4 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	delete unloading;
385			});
386	0/4 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 1 not taken. ✗ Branch 4 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 1 not taken. ✗ Branch 4 not taken.	✗	loaded->module = Traits::LoadModule(api, binary.code);
387	0/4 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	const auto [globalsAddress, globalsBytes] =
388	0/2 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 1 not taken.	✗	Traits::Symbol(api, loaded->module, "SLANG_globalParams");
389	0/4 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (globalsAddress != 0) {
390		✗	Traits::CopyToDevice(api, globalsAddress, &globals,
391	0/10 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. ✗ Branch 3 not taken. ✗ Branch 5 not taken. ✗ Branch 6 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. ✗ Branch 3 not taken. ✗ Branch 5 not taken. ✗ Branch 6 not taken.	✗	std::min<uint64_t>(globalsBytes, sizeof(globals)));
392			}
393	0/4 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	loaded->dispatchArgs =
394	0/6 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 1 not taken. ✗ Branch 3 not taken. ✗ Branch 4 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 1 not taken. ✗ Branch 3 not taken. ✗ Branch 4 not taken.	✗	Traits::Symbol(api, loaded->module, "chrono_batch_render_dispatch_args").first;
395	0/4 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	return std::shared_ptr<void>(loaded);
396		✗	}));
397	0/2 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 1 not taken.	✗	m_function = Traits::GetFunction(api, m_module->module, Name().c_str());
398
399	0/2 chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::cuda::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::cuda::Traits>) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeKernel<chrono::batch_render::device::hip::Traits>::Load(chrono::batch_render::device::KernelCode const&, chrono::batch_render::DiskCache const&, chrono::batch_render::DeviceGlobals const&)::{lambda()#1}::operator()() const::{lambda(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>)#1}::operator()(chrono::batch_render::device::compute::LoadedModule<chrono::batch_render::device::hip::Traits>) const: ✗ Branch 1 not taken.	✗	KernelStats stats = Traits::Stats(api, m_function);
400		✗	stats.cached = binary.cached;
401		✗	return stats;
402		✗	}
403
404			private:
405			using Module = Traits::Module;
406
407			Device& m_owner;
408			std::shared_ptr<LoadedModule<Traits>> m_module;
409			Traits::Function m_function{};
410			};
411
412			template <ComputeTraits Traits> class ComputeQueue : public Queue {
413			public:
414			using Device = Traits::Device;
415
416	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::ComputeQueue(chrono::batch_render::device::cuda::CudaDevice&): ✗ Branch 2 not taken. ✗ Branch 5 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::ComputeQueue(chrono::batch_render::device::hip::HipDevice&): ✗ Branch 2 not taken. ✗ Branch 5 not taken.	✗	explicit ComputeQueue(Device& device) : Queue(device) {
417	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::ComputeQueue(chrono::batch_render::device::cuda::CudaDevice&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::ComputeQueue(chrono::batch_render::device::hip::HipDevice&): ✗ Branch 1 not taken.	✗	const auto scope = Traits::Enter(device);
418	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::ComputeQueue(chrono::batch_render::device::cuda::CudaDevice&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::ComputeQueue(chrono::batch_render::device::hip::HipDevice&): ✗ Branch 1 not taken.	✗	m_stream = Traits::CreateStream(Functions());
419			try {
420	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::ComputeQueue(chrono::batch_render::device::cuda::CudaDevice&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::ComputeQueue(chrono::batch_render::device::hip::HipDevice&): ✗ Branch 1 not taken.	✗	m_pool = Traits::CreatePool(device, kPoolKeptBytes);
421	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::ComputeQueue(chrono::batch_render::device::cuda::CudaDevice&): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::ComputeQueue(chrono::batch_render::device::hip::HipDevice&): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (m_pool != Pool{}) {
422	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::ComputeQueue(chrono::batch_render::device::cuda::CudaDevice&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::ComputeQueue(chrono::batch_render::device::hip::HipDevice&): ✗ Branch 1 not taken.	✗	m_allocationStream = Traits::CreateStream(Functions());
423			}
424		✗	} catch (...) {
425	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::ComputeQueue(chrono::batch_render::device::cuda::CudaDevice&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::ComputeQueue(chrono::batch_render::device::hip::HipDevice&): ✗ Branch 1 not taken.	✗	DestroyStreams();
426		✗	throw;
427			}
428		✗	}
429
430		✗	~ComputeQueue() override {
431			try {
432	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::~ComputeQueue(): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::~ComputeQueue(): ✗ Branch 1 not taken.	✗	const auto scope = Traits::Enter(Owner());
433	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::~ComputeQueue(): ✗ Branch 1 not taken. ✗ Branch 2 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::~ComputeQueue(): ✗ Branch 1 not taken. ✗ Branch 2 not taken.	✗	for (const Pending& pending : m_pending) {
434		✗	Traits::DestroyEvent(Functions(), pending.event);
435			}
436	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::~ComputeQueue(): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::~ComputeQueue(): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	for (const Event event : m_free) {
437	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::~ComputeQueue(): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::~ComputeQueue(): ✗ Branch 1 not taken.	✗	Traits::DestroyEvent(Functions(), event);
438			}
439	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::~ComputeQueue(): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::~ComputeQueue(): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (m_anchorStream != Stream{}) {
440		✗	Traits::DestroyEvent(Functions(), m_anchor);
441		✗	Traits::DestroyStream(Functions(), m_anchorStream);
442			}
443	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::~ComputeQueue(): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::~ComputeQueue(): ✗ Branch 1 not taken.	✗	DestroyStreams();
444		✗	} catch (...) { // NOLINT(bugprone-empty-catch)
445			}
446		✗	}
447
448		✗	uint64_t Allocate(uint64_t bytes) const {
449	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Allocate(unsigned long) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Allocate(unsigned long) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (m_pool == Pool{}) {
450		✗	return Traits::Allocate(Functions(), bytes);
451			}
452			const uint64_t address =
453		✗	Traits::AllocateFrom(Functions(), m_pool, bytes, m_allocationStream);
454		✗	Traits::SynchronizeStream(Functions(), m_allocationStream);
455		✗	return address;
456			}
457
458		✗	void Free(uint64_t address) const {
459	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Free(unsigned long) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Free(unsigned long) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (m_pool == Pool{}) {
460		✗	Traits::Free(Functions(), address);
461			} else {
462		✗	Traits::FreeTo(Functions(), address, m_allocationStream);
463		✗	Traits::SynchronizeStream(Functions(), m_allocationStream);
464			}
465		✗	}
466
467			protected:
468		✗	void Execute(const CommandBuffer& queued, const CommandBuffer& commands,
469			const Timestamps* timestamps, uint64_t value) override {
470		✗	const auto scope = Traits::Enter(Owner());
471		✗	const auto* events = static_cast<const EventTimestamps*>(timestamps);
472	0/6 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 1 not taken. ✗ Branch 3 not taken. ✗ Branch 4 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 1 not taken. ✗ Branch 3 not taken. ✗ Branch 4 not taken.	✗	Run(queued, {});
473	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (!commands.Reusable()) {
474	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 1 not taken.	✗	Run(commands, {.timestamps = events});
475			} else {
476	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 1 not taken.	✗	Replay(commands, events);
477			}
478
479	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 1 not taken.	✗	const Event event = TakeEvent();
480		✗	Traits::Record(Functions(), event, m_stream, false);
481	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 1 not taken.	✗	const std::lock_guard lock(m_mutex);
482	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Execute(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::Queue::Timestamps const, unsigned long): ✗ Branch 1 not taken.	✗	m_pending.push_back({value, event});
483		✗	}
484
485		✗	std::shared_ptr<Timestamps> CreateTimestamps(uint32_t count) override {
486	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::CreateTimestamps(unsigned int): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::CreateTimestamps(unsigned int): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	return std::make_shared<EventTimestamps>(*this, count);
487			}
488
489		✗	bool SharesStreams() override {
490		✗	return true;
491			}
492
493		✗	void* NativeStream() override {
494		✗	return m_stream;
495			}
496
497		✗	void FollowStream(void* stream) override {
498		✗	const auto scope = Traits::Enter(Owner());
499		✗	const auto& api = Functions();
500	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::FollowStream(void): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::FollowStream(void): ✗ Branch 1 not taken.	✗	const Event event = Traits::CreateEvent(api, false);
501	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::FollowStream(void): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::FollowStream(void): ✗ Branch 1 not taken.	✗	Traits::Record(api, event, Traits::ApplicationStream(stream), false);
502	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::FollowStream(void): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::FollowStream(void): ✗ Branch 1 not taken.	✗	Traits::WaitEvent(api, m_stream, event);
503	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::FollowStream(void): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::FollowStream(void): ✗ Branch 1 not taken.	✗	Traits::DestroyEvent(api, event);
504		✗	}
505
506		✗	void StreamWait(void* stream, uint64_t) override {
507		✗	const auto scope = Traits::Enter(Owner());
508		✗	const auto& api = Functions();
509	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::StreamWait(void, unsigned long): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::StreamWait(void, unsigned long): ✗ Branch 1 not taken.	✗	const Event event = Traits::CreateEvent(api, false);
510	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::StreamWait(void, unsigned long): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::StreamWait(void, unsigned long): ✗ Branch 1 not taken.	✗	Traits::Record(api, event, m_stream, false);
511	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::StreamWait(void, unsigned long): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::StreamWait(void, unsigned long): ✗ Branch 1 not taken.	✗	Traits::WaitEvent(api, Traits::ApplicationStream(stream), event);
512	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::StreamWait(void, unsigned long): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::StreamWait(void, unsigned long): ✗ Branch 1 not taken.	✗	Traits::DestroyEvent(api, event);
513		✗	}
514
515		✗	uint64_t Finished() const override {
516		✗	const auto scope = Traits::Enter(Owner());
517	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Finished() const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Finished() const: ✗ Branch 1 not taken.	✗	const std::lock_guard lock(m_mutex);
518	0/10 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Finished() const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. ✗ Branch 3 not taken. ✗ Branch 5 not taken. ✗ Branch 6 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Finished() const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. ✗ Branch 3 not taken. ✗ Branch 5 not taken. ✗ Branch 6 not taken.	✗	while (!m_pending.empty() && Traits::Query(Functions(), m_pending.front().event)) {
519	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Finished() const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Finished() const: ✗ Branch 1 not taken.	✗	m_finished = m_pending.front().value;
520	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Finished() const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Finished() const: ✗ Branch 1 not taken.	✗	m_free.push_back(m_pending.front().event);
521		✗	m_pending.pop_front();
522			}
523		✗	return m_finished;
524		✗	}
525
526		✗	void WaitFor(uint64_t value) const override {
527		✗	Event event{};
528			{
529		✗	const std::lock_guard lock(m_mutex);
530	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::WaitFor(unsigned long) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::WaitFor(unsigned long) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (m_finished >= value) {
531		✗	return;
532			}
533	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::WaitFor(unsigned long) const: ✗ Branch 1 not taken. ✗ Branch 2 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::WaitFor(unsigned long) const: ✗ Branch 1 not taken. ✗ Branch 2 not taken.	✗	for (const Pending& pending : m_pending) {
534	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::WaitFor(unsigned long) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::WaitFor(unsigned long) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (pending.value >= value) {
535		✗	event = pending.event;
536		✗	break;
537			}
538			}
539		✗	}
540	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::WaitFor(unsigned long) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::WaitFor(unsigned long) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (event != Event{}) {
541		✗	const auto scope = Traits::Enter(Owner());
542	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::WaitFor(unsigned long) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::WaitFor(unsigned long) const: ✗ Branch 1 not taken.	✗	Traits::Synchronize(Functions(), event);
543		✗	}
544		✗	Finished();
545			}
546
547			private:
548			using Stream = Traits::Stream;
549			using Event = Traits::Event;
550			using Pool = Traits::Pool;
551
552			struct Pending {
553			uint64_t value;
554			Event event;
555			};
556
557		✗	const Device& Owner() const {
558	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::WaitFor(unsigned long) const: ✗ Branch 1 not taken. ✗ Branch 9 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::WaitFor(unsigned long) const: ✗ Branch 1 not taken. ✗ Branch 9 not taken.	✗	return static_cast<const Device&>(GetDevice());
559			}
560
561	0/24 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::WaitFor(unsigned long) const: ✗ Branch 1 not taken. ✗ Branch 7 not taken. ✗ Branch 10 not taken. ✗ Branch 13 not taken. ✗ Branch 16 not taken. ✗ Branch 19 not taken. ✗ Branch 22 not taken. ✗ Branch 25 not taken. ✗ Branch 28 not taken. ✗ Branch 31 not taken. ✗ Branch 37 not taken. ✗ Branch 40 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::WaitFor(unsigned long) const: ✗ Branch 1 not taken. ✗ Branch 7 not taken. ✗ Branch 10 not taken. ✗ Branch 13 not taken. ✗ Branch 16 not taken. ✗ Branch 19 not taken. ✗ Branch 22 not taken. ✗ Branch 25 not taken. ✗ Branch 28 not taken. ✗ Branch 31 not taken. ✗ Branch 37 not taken. ✗ Branch 40 not taken.	✗	const Traits::Api& Functions() const {
562	0/34 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::WaitFor(unsigned long) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. ✗ Branch 3 not taken. ✗ Branch 6 not taken. ✗ Branch 9 not taken. ✗ Branch 14 not taken. ✗ Branch 22 not taken. ✗ Branch 25 not taken. ✗ Branch 28 not taken. ✗ Branch 31 not taken. ✗ Branch 34 not taken. ✗ Branch 37 not taken. ✗ Branch 40 not taken. ✗ Branch 43 not taken. ✗ Branch 46 not taken. ✗ Branch 56 not taken. ✗ Branch 59 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::WaitFor(unsigned long) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. ✗ Branch 3 not taken. ✗ Branch 6 not taken. ✗ Branch 9 not taken. ✗ Branch 14 not taken. ✗ Branch 22 not taken. ✗ Branch 25 not taken. ✗ Branch 28 not taken. ✗ Branch 31 not taken. ✗ Branch 34 not taken. ✗ Branch 37 not taken. ✗ Branch 40 not taken. ✗ Branch 43 not taken. ✗ Branch 46 not taken. ✗ Branch 56 not taken. ✗ Branch 59 not taken.	✗	return Owner().Functions();
563			}
564
565			struct EventTimestamps : Timestamps {
566	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps::EventTimestamps(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits> const&, unsigned long): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps::EventTimestamps(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits> const&, unsigned long): ✗ Branch 1 not taken.	✗	EventTimestamps(const ComputeQueue& owner, size_t count) : queue(owner) {
567	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps::EventTimestamps(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits> const&, unsigned long): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps::EventTimestamps(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits> const&, unsigned long): ✗ Branch 1 not taken.	✗	const auto scope = Traits::Enter(owner.Owner());
568	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps::EventTimestamps(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits> const&, unsigned long): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps::EventTimestamps(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits> const&, unsigned long): ✗ Branch 1 not taken.	✗	events.reserve(count);
569	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps::EventTimestamps(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits> const&, unsigned long): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps::EventTimestamps(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits> const&, unsigned long): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	for (size_t i = 0; i < count; ++i) {
570	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps::EventTimestamps(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits> const&, unsigned long): ✗ Branch 1 not taken. ✗ Branch 4 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps::EventTimestamps(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits> const&, unsigned long): ✗ Branch 1 not taken. ✗ Branch 4 not taken.	✗	events.push_back(Traits::CreateEvent(owner.Functions(), true));
571			}
572		✗	}
573
574		✗	~EventTimestamps() override {
575			try {
576		✗	const auto scope = Traits::Enter(queue.Owner());
577	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps::~EventTimestamps(): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps::~EventTimestamps(): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	for (const Event event : events) {
578	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps::~EventTimestamps(): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps::~EventTimestamps(): ✗ Branch 1 not taken.	✗	Traits::DestroyEvent(queue.Functions(), event);
579			}
580		✗	} catch (...) { // NOLINT(bugprone-empty-catch)
581			}
582		✗	}
583
584			EventTimestamps(const EventTimestamps&) = delete;
585			EventTimestamps& operator=(const EventTimestamps&) = delete;
586
587		✗	bool Read(std::span<uint64_t> ns) const override {
588		✗	const auto scope = Traits::Enter(queue.Owner());
589	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps::Read(std::span<unsigned long, 18446744073709551615ul>) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps::Read(std::span<unsigned long, 18446744073709551615ul>) const: ✗ Branch 1 not taken.	✗	const auto [anchor, anchorNs] = queue.Anchor();
590	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps::Read(std::span<unsigned long, 18446744073709551615ul>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps::Read(std::span<unsigned long, 18446744073709551615ul>) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	for (size_t i = 0; i < ns.size(); ++i) {
591	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps::Read(std::span<unsigned long, 18446744073709551615ul>) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps::Read(std::span<unsigned long, 18446744073709551615ul>) const: ✗ Branch 1 not taken.	✗	ns[i] = anchorNs - Traits::ElapsedNs(queue.Functions(), events[i], anchor);
592			}
593		✗	return true;
594		✗	}
595
596			const ComputeQueue& queue;
597			std::vector<Event> events;
598			};
599
600		✗	struct Graph {
601			Traits::Graph graph{};
602			uint64_t data = 0;
603			std::vector<std::pair<const ComputeKernel<Traits>*, uint64_t>> counts;
604			};
605
606			struct RunDesc {
607			uint64_t data = 0;
608			const EventTimestamps* timestamps = nullptr;
609			bool capturing = false;
610			};
611
612		✗	void Run(const CommandBuffer& list, const RunDesc& desc) const {
613		✗	const auto& api = Functions();
614	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Run(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::RunDesc const&) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Run(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::RunDesc const&) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	for (const Command& command : list.Commands()) {
615		✗	std::visit(
616			Overloaded{
617		✗	[&](const CopyCommand& c) {
618		✗	Traits::Copy(api, c.dst->Address() + c.dstOffset,
619		✗	c.src->Address() + c.srcOffset, c.bytes, m_stream);
620			},
621		✗	[&](const FillCommand& c) {
622		✗	Traits::Fill(api, c.dst->Address() + c.offset, c.value, c.bytes, m_stream);
623			},
624		✗	[&](const UpdateCommand& c) {
625	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Run(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::RunDesc const&) const::{lambda(chrono::batch_render::device::UpdateCommand const&)#1}::operator()(chrono::batch_render::device::UpdateCommand const&) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Run(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::RunDesc const&) const::{lambda(chrono::batch_render::device::UpdateCommand const&)#1}::operator()(chrono::batch_render::device::UpdateCommand const&) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (desc.data != 0) {
626		✗	Traits::Copy(api, c.dst->Address() + c.offset, desc.data + c.data,
627		✗	c.bytes, m_stream);
628			} else {
629		✗	Traits::Upload(api, c.dst->Address() + c.offset, list.Data(c), c.bytes,
630		✗	m_stream);
631			}
632			},
633			[&](const BarrierCommand&) {},
634		✗	[&](const ZoneCommand& c) {
635	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Run(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::RunDesc const&) const::{lambda(chrono::batch_render::device::ZoneCommand const&)#1}::operator()(chrono::batch_render::device::ZoneCommand const&) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Run(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::RunDesc const&) const::{lambda(chrono::batch_render::device::ZoneCommand const&)#1}::operator()(chrono::batch_render::device::ZoneCommand const&) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (desc.timestamps != nullptr) {
636		✗	Traits::Record(api,
637		✗	desc.timestamps->events[2 * c.index + (c.begin ? 0 : 1)],
638	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Run(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::RunDesc const&) const::{lambda(chrono::batch_render::device::ZoneCommand const&)#1}::operator()(chrono::batch_render::device::ZoneCommand const&) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Run(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::RunDesc const&) const::{lambda(chrono::batch_render::device::ZoneCommand const&)#1}::operator()(chrono::batch_render::device::ZoneCommand const&) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	m_stream, desc.capturing);
639			}
640			},
641			[&](const BuildAccelCommand&) {},
642		✗	[&](const TextureCopyCommand& c) {
643		✗	static_cast<const ComputeTexture<Traits>&>(*c.dst).Copy(
644		✗	c.level, c.src->Address() + c.srcOffset, m_stream);
645			},
646		✗	[&](const DispatchCommand& c) {
647		✗	const uint64_t counts =
648	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Run(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::RunDesc const&) const::{lambda(chrono::batch_render::device::DispatchCommand const&)#1}::operator()(chrono::batch_render::device::DispatchCommand const&) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Run(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::RunDesc const&) const::{lambda(chrono::batch_render::device::DispatchCommand const&)#1}::operator()(chrono::batch_render::device::DispatchCommand const&) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	c.counts != nullptr ? c.counts->Address() + c.countsOffset : 0;
649		✗	Kernel(c).Launch(api, c.groups, counts, list.Data(c), c.paramBytes,
650		✗	m_stream);
651			},
652			},
653			command);
654			}
655		✗	}
656
657		✗	static const ComputeKernel<Traits>& Kernel(const DispatchCommand& dispatch) {
658		✗	return static_cast<const ComputeKernel<Traits>&>(*dispatch.kernel);
659			}
660
661	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Replay(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Replay(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	void Replay(const CommandBuffer& commands, const EventTimestamps* timestamps) const {
662		✗	std::shared_ptr<void>& recording = Recording(commands);
663	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Replay(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Replay(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (!recording) {
664	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Replay(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const: ✗ Branch 1 not taken. ✗ Branch 2 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Replay(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const: ✗ Branch 1 not taken. ✗ Branch 2 not taken.	✗	recording = Capture(commands, timestamps);
665			}
666		✗	const Graph& graph = static_cast<const Graph>(recording.get());
667		✗	Traits::LaunchGraph(Functions(), graph.graph, m_stream);
668	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Replay(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Replay(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	for (const auto& [kernel, counts] : graph.counts) {
669		✗	kernel->Counts() = counts;
670			}
671		✗	}
672
673		✗	Event TakeEvent() {
674			{
675		✗	const std::lock_guard lock(m_mutex);
676	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::TakeEvent(): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::TakeEvent(): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (!m_free.empty()) {
677		✗	const Event event = m_free.back();
678		✗	m_free.pop_back();
679		✗	return event;
680			}
681		✗	}
682		✗	return Traits::CreateEvent(Functions(), false);
683			}
684
685		✗	std::shared_ptr<Graph> Capture(const CommandBuffer& commands,
686			const EventTimestamps* timestamps) const {
687		✗	const auto& api = Functions();
688		✗	std::shared_ptr<Graph> graph(new Graph, [this](Graph* freed) {
689			try {
690	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 1 not taken.	✗	const auto scope = Traits::Enter(Owner());
691	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (freed->graph != typename Traits::Graph{}) {
692	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 1 not taken.	✗	Traits::DestroyGraph(Functions(), freed->graph);
693			}
694	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (freed->data != 0) {
695	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 1 not taken.	✗	Free(freed->data);
696			}
697		✗	} catch (...) { // NOLINT(bugprone-empty-catch)
698			}
699	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	delete freed;
700			});
701
702		✗	const std::span<const std::byte> data = commands.Data();
703	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	const bool updates = std::ranges::any_of(commands.Commands(), [](const Command& command) {
704	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	return std::holds_alternative<UpdateCommand>(command);
705			});
706	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (updates) {
707	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 1 not taken. ✗ Branch 4 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 1 not taken. ✗ Branch 4 not taken.	✗	graph->data = Allocate(data.size());
708	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 1 not taken.	✗	Traits::CopyToDevice(api, graph->data, data.data(), data.size());
709			}
710
711	0/8 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. ✗ Branch 2 not taken. ✗ Branch 3 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. ✗ Branch 2 not taken. ✗ Branch 3 not taken.	✗	for (const Command& command : commands.Commands()) {
712		✗	if (const auto* dispatch = std::get_if<DispatchCommand>(&command)) {
713		✗	Kernel(*dispatch).Counts() = ComputeKernel<Traits>::kUnknownCounts;
714			}
715			}
716	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 1 not taken. ✗ Branch 4 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 1 not taken. ✗ Branch 4 not taken.	✗	graph->graph = RecordGraph(commands, graph->data, timestamps);
717	0/6 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Graph) const: ✗ Branch 1 not taken. ✗ Branch 5 not taken. ✗ Branch 6 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Capture(chrono::batch_render::device::CommandBuffer const&, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const::{lambda(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph)#1}::operator()(chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Graph) const: ✗ Branch 1 not taken. ✗ Branch 5 not taken. ✗ Branch 6 not taken.	✗	graph->counts = KernelCounts(commands);
718		✗	return graph;
719		✗	}
720
721		✗	Traits::Graph RecordGraph(const CommandBuffer& commands, uint64_t data,
722			const EventTimestamps* timestamps) const {
723		✗	const auto& api = Functions();
724		✗	Traits::BeginCapture(api, m_stream);
725			try {
726	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::RecordGraph(chrono::batch_render::device::CommandBuffer const&, unsigned long, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const: ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::RecordGraph(chrono::batch_render::device::CommandBuffer const&, unsigned long, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const: ✗ Branch 1 not taken.	✗	Run(commands, {.data = data, .timestamps = timestamps, .capturing = true});
727		✗	} catch (...) {
728			try {
729	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::RecordGraph(chrono::batch_render::device::CommandBuffer const&, unsigned long, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const: ✗ Branch 1 not taken. ✗ Branch 4 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::RecordGraph(chrono::batch_render::device::CommandBuffer const&, unsigned long, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const: ✗ Branch 1 not taken. ✗ Branch 4 not taken.	✗	Traits::DestroyGraph(api, Traits::EndCapture(api, m_stream));
730	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::RecordGraph(chrono::batch_render::device::CommandBuffer const&, unsigned long, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::EventTimestamps const) const: ✗ Branch 2 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::RecordGraph(chrono::batch_render::device::CommandBuffer const&, unsigned long, chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::EventTimestamps const) const: ✗ Branch 2 not taken.	✗	} catch (...) { // NOLINT(bugprone-empty-catch)
731			}
732		✗	throw;
733			}
734		✗	return Traits::EndCapture(api, m_stream);
735			}
736
737			static std::vector<std::pair<const ComputeKernel<Traits>*, uint64_t>>
738		✗	KernelCounts(const CommandBuffer& commands) {
739		✗	std::vector<std::pair<const ComputeKernel<Traits>*, uint64_t>> counts;
740	0/8 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::KernelCounts(chrono::batch_render::device::CommandBuffer const&): ✗ Branch 0 not taken. ✗ Branch 1 not taken. ✗ Branch 2 not taken. ✗ Branch 3 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::KernelCounts(chrono::batch_render::device::CommandBuffer const&): ✗ Branch 0 not taken. ✗ Branch 1 not taken. ✗ Branch 2 not taken. ✗ Branch 3 not taken.	✗	for (const Command& command : commands.Commands()) {
741		✗	if (const auto* dispatch = std::get_if<DispatchCommand>(&command)) {
742		✗	const ComputeKernel<Traits>* kernel = &Kernel(*dispatch);
743	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::KernelCounts(chrono::batch_render::device::CommandBuffer const&): ✗ Branch 1 not taken. ✗ Branch 2 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::KernelCounts(chrono::batch_render::device::CommandBuffer const&): ✗ Branch 1 not taken. ✗ Branch 2 not taken.	✗	if (std::ranges::none_of(counts,
744	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::KernelCounts(chrono::batch_render::device::CommandBuffer const&): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::KernelCounts(chrono::batch_render::device::CommandBuffer const&): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	[&](const auto& held) { return held.first == kernel; })) {
745	0/2 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::KernelCounts(chrono::batch_render::device::CommandBuffer const&): ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::KernelCounts(chrono::batch_render::device::CommandBuffer const&): ✗ Branch 1 not taken.	✗	counts.emplace_back(kernel, kernel->Counts());
746			}
747			}
748			}
749		✗	return counts;
750		✗	}
751
752	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Anchor() const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Anchor() const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	std::pair<Event, uint64_t> Anchor() const {
753		✗	const auto& api = Functions();
754	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::Anchor() const: ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::Anchor() const: ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (m_anchorStream == Stream{}) {
755		✗	m_anchorStream = Traits::CreateStream(api);
756		✗	m_anchor = Traits::CreateEvent(api, true);
757			}
758		✗	const uint64_t before = profiler::Now();
759		✗	Traits::Record(api, m_anchor, m_anchorStream, false);
760		✗	Traits::Synchronize(api, m_anchor);
761		✗	const uint64_t after = profiler::Now();
762		✗	return {m_anchor, before + (after - before) / 2};
763			}
764
765		✗	void DestroyStreams() {
766	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::DestroyStreams(): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::DestroyStreams(): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (m_allocationStream != Stream{}) {
767		✗	Traits::SynchronizeStream(Functions(), m_allocationStream);
768		✗	Traits::DestroyStream(Functions(), m_allocationStream);
769			}
770	0/4 chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::cuda::Traits>::DestroyStreams(): ✗ Branch 0 not taken. ✗ Branch 1 not taken. chrono::batch_render::device::compute::ComputeQueue<chrono::batch_render::device::hip::Traits>::DestroyStreams(): ✗ Branch 0 not taken. ✗ Branch 1 not taken.	✗	if (m_pool != Pool{}) {
771		✗	Traits::DestroyPool(Functions(), m_pool);
772			}
773		✗	Traits::DestroyStream(Functions(), m_stream);
774		✗	}
775
776			Stream m_stream{};
777			Pool m_pool{};
778			Stream m_allocationStream{};
779			mutable Stream m_anchorStream{};
780			mutable Event m_anchor{};
781			mutable std::mutex m_mutex;
782			mutable std::deque<Pending> m_pending;
783			mutable std::vector<Event> m_free;
784			mutable uint64_t m_finished = 0;
785			};
786
787			}
788