いまどきのVulkan

Vulkan
Modern Vulkan
NAOMASA MATSUBAYASHI
Twitter: @fadis_
͍·Ͳ͖ͷ
ιʔείʔυ: https://github.com/Fadis/gct/tree/kernelvm-online-4

View Slide

Vulkan
GPUΛૢ࡞͢Δҝͷ
ΫϩεϓϥοτϑΥʔϜͳAPI
https://www.vulkan.org/

View Slide

Vulkan
GPUΛૢ࡞͢Δҝͷ
ΫϩεϓϥοτϑΥʔϜͳAPI
https://www.vulkan.org/
Windows
Nintendo Switch
Stadia
Android
Linux
MoltenVK(macOS iOS iPadOS)
͋ͱFuchsia΍QNX΋ରԠͯ͠Δ

View Slide

GPU
3DάϥϑΟΫεΛඳ͘ҝͷઐ༻ϋʔυ΢ΣΞ
ϑϨʔϜόοϑΝͷ಺༰Λը໘ʹૹΔػߏ
+
20ੈلͷ

View Slide

3DάϥϑΟΫεΛඳ͘ҝͷઐ༻ϋʔυ΢ΣΞ
+
GPU
3DάϥϑΟΫεʹ
ཁٻ͞ΕΔܭࢉ͕ෳࡶʹͳͬͯ
͋ͬͱ͍͏ؒʹഁ୼

View Slide

GPU
೚ҙͷܭࢉΛߦ͏ϓϩηοα
+
+
࣮ߦՄೳόΠφϦͱσʔλΛஔ͍͓ͯ͘ϝϞϦ
21ੈلͷ

View Slide

GPU
+
+
21ੈلͷ
Ͳ͏ͯ͜͠ͷํ๏Ͱ
CPUΑΓߴ଎ʹܭࢉͰ͖Δͷ?

View Slide

GPU
+
+
21ੈلͷ
େྔͷ

View Slide

float x32
Tensor
Core
ϩʔυετΞ
σΟεύον໋ྩΩϟογϡ
ϨδελόϯΫ
GeForce
RTX3080ͷ৔߹
ALU
εʔύʔεΧϥͷҝͷ
ෳࡶͳґଘؔ܎ͷ
νΣοΫ౳͸࣋ͨͳ͍
∴͜ͷϓϩηοα1ݸͷ
τϥϯδελ਺͸
খ͘͞཈͑ΒΕΔ
Warp
(Subgroup)

View Slide

float x128
ڞ༗ϝϞϦ L1Ωϟογϡ
RT Core
GeForce
RTX3080ͷ৔߹
Streaming
Multiprocessor
(Work Group)

View Slide

float x256
GeForce
RTX3080ͷ৔߹
Texture
Processing
Cluster
PolyMorph

View Slide

float x1536
ϥελϥΠβ
Raster Operators
Graphics
Processing
Clusters

View Slide

float x10752
PCI-ExpressϗετΠϯλʔϑΣʔε
NVLinkϗετΠϯλʔϑΣʔε
L2Ωϟογϡ
Graphics
Processing
Unit
(Physical Device)

View Slide

float x 21504
PCI-Express
NVLink
Device
Group

View Slide

1ΫϩοΫͰେྔͷσʔλʹରͯ͠ԋࢉ
ݸʑͷϓϩηοα͕গʑ஗ͯ͘΋CPUΛѹ౗Ͱ͖Δ
ԿͰCPU͸ͦ͏͠ͳ͍ͷ?
CPUͷxxഒ଎͍
·͔͡Α

View Slide

1ΫϩοΫͰܭࢉͰ͖Δ਺Ҏ্ͷσʔλ͕ಉ࣌ʹແ͍ͱ
Կ΋͠ͳ͍ԋࢉث͕ੜ͡
ͨͩͷ஗͍ܭࢉػʹͳΔ
஋1
஋2
஋3
࢖ΘΕͳ͍ԋࢉث
શ෦Ͱ3ݸͷ
σʔλ
͜ͷ৚݅ΛຬͨͤΔ͔Ͳ͏͔͸λεΫʹґΔ

View Slide

े෼ͳฒྻ౓
͕͋Δ
৽छͷλεΫ
Yes
No

View Slide

෼ۀ
OSͱ͔໘౗ͳͷ͸೚ͤͨ
Զ͸σΟʔϓϥʔχϯάͱ͔͚ͩ͢Δ
ͻͰ͐

View Slide

GPUͷಈ͔͠ํ
1. GPUͷϝϞϦʹσʔλΛૹΔ
2. GPU্Ͱ࣮ߦՄೳόΠφϦΛ࣮ߦ͢Δ
3. GPUͷϝϞϦ͔Β݁ՌΛऔΓग़͢
͍ΖΜͳϕϯμʔͷGPU͕͋Δ͚Ͳ
ϕϯμʔʹґΒͣ͜ͷૢ࡞Λ͢ΔAPI͕Vulkan
ۃΊͯࡶͳ
ೖྗ
ೖྗ ग़ྗ
ग़ྗ

View Slide

GPUͷϝϞϦʹσʔλΛૹΔ
MMU
ී௨ʹmallocͨ͠ϝϞϦ͸
PCI-ExpressͷσόΠε͔Β͸
࿈ଓͨ͠ྖҬʹݟ͑ͳ͍
ҟͳΔMMUΛհͯ͠
ϝϞϦΛݟ͍ͯΔ
0x4000
0x4000
IOMMU
0x4000ͷσʔλΛ͍࣋ͬͯͬͯΑ

View Slide

MMU͔Β΋IOMMU͔Β΋
ಉ͡Α͏ʹݟ͑ΔྖҬΛ
ϝΠϯϝϞϦʹ֬อ͢Δ
0x4000
0x1000
IOMMU
0x1000
GPUʹૹΓ͍ͨσʔλΛ
͜ͷྖҬʹίϐʔ͢Δ
MMU

View Slide

CPU͕ॻ͖׵͑Δ͔΋͠Εͳ͍ϝϞϦΛ
GPU͸ΩϟογϡͰ͖ͳ͍
0x1000
IOMMU
0x5000
CPUͷϝϞϦ্ͷྖҬͷσʔλΛ
GPUͷϝϞϦ্ʹ֬อͨ͠ྖҬʹ
ίϐʔ͢Δ
CPUͷϝϞϦ
GPUͷϝϞϦ

View Slide

0x1000
IOMMU
0x5000
MMU
0x4000
0x1000
͜ͷίϐʔ͸memcpyͰྑ͍
͜ͷྖҬͷ֬อ͸
mallocͰྑ͍
͜ͷྖҬͷ֬อʹ͸
ઐ༻ͷAPI͕ཁΔ
͜ͷྖҬͷ֬อʹ΋
ઐ༻ͷAPI͕ཁΔ
͜ͷίϐʔΛߦ͏ʹ͸
ઐ༻ͷAPI͕ཁΔ

View Slide

0x1000
IOMMU
0x5000
MMU
0x4000
0x1000
͜ͷίϐʔ͸memcpyͰྑ͍
͜ͷྖҬͷ֬อ͸
mallocͰྑ͍
vkAllocateMemory
vkCmdCopyBuffer vkAllocateMemory

View Slide

0x1000
IOMMU
0x5000
MMU
0x4000
0x1000
͜͏͍͏
ྖҬͷ͜ͱΛ
Staging Buffer
ͱݺͿ

View Slide

GPUͷϝϞϦ͔Β݁ՌΛऔΓग़͢
0x1000
IOMMU
0x5000
MMU
0x4000
0x1000
vkAllocateMemory
vkCmdCopyBuffer vkAllocateMemory
memcpy
malloc
CPUʹσʔλΛฦ࣌͢΋ಉ͡ํ๏Ͱ

View Slide

0x1000
IOMMU
0x5000
MMU
0x4000
0x1000
CPU͔Βίϐʔͨ͠
ූ߸෇͖੔਺΍ුಈখ਺఺਺Λ
GPU͸ม׵ͳ͠Ͱ
ಉ͡Α͏ʹղऍͰ͖ͳ͚Ε͹ͳΒͳ͍

View Slide

https://www.khronos.org/registry/vulkan/specs/1.0/html/chap3.html#fundamentals-host-environment
https://www.khronos.org/registry/vulkan/specs/1.0/html/chap36.html#spirvenv-precision-operation
32͓Αͼ64bitͷුಈখ਺఺਺͸IEEE Std 754-2008
ූ߸෇͖੔਺͸2ͷิ਺දݱ
ΤϯσΟΞϯ͸CPUͱGPUͰಉ͡΋ͷΛαϙʔτ
NaN
NaN
Vulkan 1.0ͷن֨ΑΓ
VulkanରԠ؀ڥͷCPUͱGPUͨΔ΋ͷ
͜͏ܾ·͍ͬͯΔͷͰ
ͦͷ··ίϐʔͨ͠஋͕ಡΊΔ

View Slide

"memory_props": { "basic": {
"memoryHeaps": [
{ "flags": 1, "size": 8589934592 },
{ "flags": 0, "size": 12528737280 },
{ "flags": 1, "size": 257949696 }
],
"memoryTypes": [
{ "heapIndex": 1, "propertyFlags": 0 },
{ "heapIndex": 2, "propertyFlags": 7 }
]
}}
vkGetPhysicalDeviceMemoryPropertiesͰ࢖͑ΔϝϞϦΛௐ΂Δ
GPUͷϝϞϦʹ
ಠཱͨ͠ώʔϓ͕2ͭ
CPUͷϝϞϦʹ
ಠཱͨ͠ώʔϓ͕1ͭ

View Slide

"memoryHeaps": [
{ "flags": 1, "size": 8589934592 },
{ "flags": 0, "size": 12528737280 },
{ "flags": 1, "size": 257949696 }
],
"memoryTypes": [
]
}}
͜ͷล͸
ಛघ༻్ͳͷͰ
ࠓ͸ແࢹ
ϝϞϦλΠϓ
ͲΜͳৼΔ෣͍Λ͢Δ
ϝϞϦΛ֬อͰ͖Δ͔

View Slide

"memoryHeaps": [
{ "flags": 1, "size": 8589934592 },
{ "flags": 0, "size": 12528737280 },
{ "flags": 1, "size": 257949696 }
],
"memoryTypes": [
]
}}
GPUͷϝϞϦʹ
GPUͷΈ͔Βݟ͑ΔྖҬΛ
֬อͰ͖Δ
CPUͷϝϞϦʹCPU͔Βݟ͑ͯ
CPU͕Ωϟογϡ͠ͳ͍ྖҬΛ
֬อͰ͖Δ
CPU͕Ωϟογϡ͢ΔྖҬΛ
֬อͰ͖Δ
GPUͷϝϞϦʹCPU͔Βݟ͑ͯ
֬อͰ͖Δ

View Slide

ಛघͳϝϞϦ͸vkAllocateMemoryͰ֬อ
VkResult vkAllocateMemory(
VkDevice device,
const VkMemoryAllocateInfo* pAllocateInfo,
const VkAllocationCallbacks* pAllocator,
VkDeviceMemory* pMemory
);
typedef struct VkMemoryAllocateInfo {
VkStructureType sType;
const void* pNext;
VkDeviceSize allocationSize;
uint32_t memoryTypeIndex;
} VkMemoryAllocateInfo;
͜ͷαΠζ
͜ͷϝϞϦλΠϓͷϝϞϦΛ
͘Ε
͜ͷGPU༻ʹ

View Slide

֬อͨ͠ϝϞϦΛ
ܭࢉʹ࢖͏σʔλΛஔ͘
όοϑΝͱͯ͠࢖͏
ͱ͍͏ҙࢥදࣔΛ͢Δ
VkResult vkCreateBuffer(
VkDevice device,
const VkBufferCreateInfo* pCreateInfo,
VkBuffer* pBuffer
); typedef struct VkBufferCreateInfo {
const void* pNext;
VkBufferCreateFlags flags;
VkDeviceSize size;
VkBufferUsageFlags usage;
VkSharingMode sharingMode;
uint32_t queueFamilyIndexCount;
const uint32_t* pQueueFamilyIndices;
} VkBufferCreateInfo;
͜ͷαΠζͷ
͜ͷGPU༻ʹ
͜Μͳ༻్ͷόοϑΝΛ
࡞ͬͯ
VkDeviceMemory VkBuffer
ϝϞϦͷத਎͸൚༻తͳσʔλͰ͢

View Slide

֬อͨ͠ϝϞϦΛ
ܭࢉʹ࢖͏σʔλΛஔ͘
όοϑΝͱͯ͠࢖͏
ͱ͍͏ҙࢥදࣔΛ͢Δ
VkResult vkCreateBuffer(
VkDevice device,
const VkBufferCreateInfo* pCreateInfo,
VkBuffer* pBuffer
); typedef struct VkBufferCreateInfo {
const void* pNext;
VkBufferCreateFlags flags;
VkDeviceSize size;
VkBufferUsageFlags usage;
} VkBufferCreateInfo;
͜ͷαΠζͷ
͜ͷGPU༻ʹ
VkResult vkBindBufferMemory(
VkDevice device,
VkBuffer buffer,
VkDeviceMemory memory,
VkDeviceSize memoryOffset
);
͜ͷϝϞϦΛ
࢖͏
͜ͷόοϑΝ͸
͜Μͳ༻్ͷόοϑΝΛ
࡞ͬͯ

View Slide

"memoryHeaps": [
{ "flags": 1, "size": 8589934592 },
{ "flags": 0, "size": 12528737280 },
{ "flags": 1, "size": 257949696 }
],
"memoryTypes": [
]
}}
CPU͔Βݟ͑Δଐੑͷ͍ͭͨϝϞϦ͸
GPUͷϝϞϦʹ
GPUͷΈ͔Βݟ͑ΔྖҬΛ
֬อͰ͖Δ
֬อͰ͖Δ
֬อͰ͖Δ
֬อͰ͖Δ

View Slide

]
}}
֬อͰ͖Δ
֬อͰ͖Δ
VkResult vkMapMemory(
VkDevice device,
VkDeviceMemory memory,
VkDeviceSize offset,
VkDeviceSize size,
VkMemoryMapFlags flags,
void** ppData
);
͜ͷϝϞϦͷ
ઌ಄ΞυϨε͕ฦͬͯ͘Δ
vkMapMemory͔ͯ͠ΒvkUnmapMemory͢Δ·Ͱͷؒ
ϓϩηεͷΞυϨεۭؒʹϝϞϦ͕Ϛοϓ͞ΕΔ
͜ͷҐஔ͔Β
͜ͷ௕͞ͷൣғͷ

View Slide

ίϚϯυ
ίϚϯυ
݁Ռ
݁Ռ
GPUʹԿ͔Λͤ͞Δʹ͸
ΩϡʔʹίϚϯυΛྲྀ͢
vkCmdCopyBufferͰ
CPUͷϝϞϦʹ͋ΔσʔλΛ
GPUʹҾͬுΒ͍ͤͨ

View Slide

ίϚϯυόοϑΝ
ίϚϯυ
ίϚϯυ
ίϚϯυ͸
ίϚϯυόοϑΝʹଋͶͯૹΔ
ίϚϯυόοϑΝͷ
಺༰͕׬ྃͨ͠
ίϚϯυόοϑΝ1ͭʹରͯ͠
࣮ߦ׬ྃ௨஌͕1ͭฦͬͯ͘Δ

View Slide

1ͭͷGPU͕
ෳ਺ͷΩϡʔΛ͍࣋ͬͯΔࣄ͕͋Δ
ಉҰͷΩϡʔʹର͢Δॻ͖ࠐΈ͸
ഉଞతʹߦ͏ඞཁ͕͋Δ͕
ҟͳΔΩϡʔʹର͢Δॻ͖ࠐΈ͸
ෳ਺ͷCPU͔Βಉ࣌ʹߦΘΕͯ΋ྑ͍

View Slide

"queue_family": [
{
"basic": {
"minImageTransferGranularity": {
...
},
"queueCount": 16,
"queueFlags": 15,
"timestampValidBits": 64
}
},
{
"basic": {
...
},
"queueCount": 2,
"queueFlags": 12,
}
vkGetPhysicalDeviceQueueFamilyPropertiesͰ࢖͑ΔΩϡʔΛௐ΂Δ
άϥϑΟοΫʹؔΘΔίϚϯυΛྲྀͤΔ
GPUͰܭࢉ͢ΔҝͷίϚϯυΛྲྀͤΔ
σʔλͷసૹͷҝͷίϚϯυΛྲྀͤΔ
͜͏͍͏Ωϡʔ͕16ຊ
͜͏͍͏Ωϡʔ͕2ຊ

View Slide

}
},
{
"basic": {
...
},
"queueCount": 2,
"queueFlags": 12,
}
},
{
"basic": {
...
},
"queueCount": 8,
"queueFlags": 14,
}
},
GPUͷԋࢉثͱ͸ಠཱʹಈ͚ΔDMA͕
8ج͋Δͱ͍͏͜ͱ

View Slide

ίϚϯυϓʔϧ
ίϚϯυόοϑΝ ίϚϯυόοϑΝ
⋯
ίϚϯυόοϑΝ
ίϚϯυ
vkAllocateCommandBuffers
ίϚϯυ͸ઐ༻ͷϝϞϦʹ
ੵ·ͳ͚Ε͹ͳΒͳ͍ࣄ͕͋ΔͷͰ
ઐ༻ͷϝϞϦϓʔϧ͔ΒׂΓ౰ͯ
vkCreateCommandPool
σόΠε
ϓʔϧΛ࡞੒
ίϚϯυόοϑΝΛऔಘ
vkFreeCommandBuffers ίϚϯυόοϑΝΛฦ٫
࢖͍ऴΘͬͨΒ

View Slide

ίϚϯυϓʔϧ
ίϚϯυόοϑΝ ίϚϯυόοϑΝ
⋯
ίϚϯυόοϑΝ
vkCmdCopyBuffer
vkAllocateCommandBuffers
vkCreateCommandPool
vkCmdCopyBufferΛ
ίϚϯυόοϑΝʹੵΜͰ
ΩϡʔʹSubmit࣮ͯ͠ߦ
VkResult vkQueueSubmit(
VkQueue queue,
uint32_t submitCount,
const VkSubmitInfo* pSubmits,
VkFence fence
);
͜ͷΩϡʔʹ

View Slide

vkCmdCopyBuffer
ίϚϯυόοϑΝʹੵΜͰ
ΩϡʔʹSubmit࣮ͯ͠ߦ
VkQueue queue,
VkFence fence
);
͜ͷΩϡʔʹ
typedef struct VkSubmitInfo {
const void* pNext;
uint32_t waitSemaphoreCount;
const VkSemaphore* pWaitSemaphores;
const VkPipelineStageFlags* pWaitDstStageMask;
uint32_t commandBufferCount;
const VkCommandBuffer* pCommandBuffers;
uint32_t signalSemaphoreCount;
const VkSemaphore* pSignalSemaphores;
} VkSubmitInfo;
͜ͷ
ίϚϯυόοϑΝΛ
ྲྀͯ͠

View Slide

VkQueue queue,
VkFence fence
);
VkResult vkWaitForFences(
VkDevice device,
uint32_t fenceCount,
const VkFence* pFences,
VkBool32 waitAll,
uint64_t timeout
);
͜͜ͰSubmitͨ͠
ίϚϯυόοϑΝͷ
಺༰͕
׬ྃ͢Δ͔
timeoutͷ࣌ؒܦա͢Δ·Ͱ
଴ػͯ͠
VkResult vkCreateFence(
VkDevice device,
const VkFenceCreateInfo* pCreateInfo,
VkFence* pFence
);
FenceΛ࡞ͬͯ׬ྃ௨஌Λड͚औΔ

View Slide

GPUͷಈ͔͠ํ
1. GPUͷϝϞϦʹσʔλΛૹΔ
2. GPU্Ͱ࣮ߦՄೳόΠφϦΛ࣮ߦ͢Δ
3. GPUͷϝϞϦ͔Β݁ՌΛऔΓग़͢
ۃΊͯࡶͳ
ೖྗ
ೖྗ ग़ྗ
ग़ྗ
γΣʔμ

View Slide

GeForceͯ͞͠ಈ͘ͳΒ
RADEONͯ͞͠΋ಈ͘΍Ζ
PCࣗ࡞erͷҰൠతͳࢥߟ
GPUͷ໋ྩηοτ͸ϕϯμʔຖʹҟͳΔ
͕ɺͳ͔ͳ͔ཧղͯ͠΋Β͑ͳ͍

View Slide

--- gcn.list 2021-11-09 02:04:47.899271324 +0900
+++ rdna2.list 2021-11-09 02:22:47.976688357 +0900
@@ -1,29 +1,41 @@
-V_ADDC_U32
+V_ADD3_U32
+V_ADD_CO_CI_U32
+V_ADD_CO_U32
+V_ADD_F16
V_ADD_F32
V_ADD_F64
-V_ADD_I32
+V_ADD_LSHL_U32
+V_ADD_NC_I16
+V_ADD_NC_I32
+V_ADD_NC_U16
+V_ADD_NC_U32
V_ALIGNBIT_B32
V_ALIGNBYTE_B32
V_AND_B32
-V_ASHRREV_I32
-V_ASHR_I32
-V_ASHR_I64
+V_AND_OR_B32
+V_ASHRREV_B32
+V_ASHRREV_I16
+V_ASHRREV_I64
V_BCNT_U32_B32
V_BFE_I32
V_BFE_U32
V_BFI_B32
V_BFM_B32
V_BFREV_B32
+V_CEIL_F16
V_CEIL_F32
V_CEIL_F64
V_CLREXCP
V_CNDMASK_B32
+V_COS_F16
V_COS_F32
V_CUBEID_F32
V_CUBEMA_F32
V_CUBESC_F32
V_CUBETC_F32
V_CVT_F16_F32
+V_CVT_F16_I16
+V_CVT_F16_U16
V_CVT_F32_F16
V_CVT_F32_F64
V_CVT_F32_I32
@@ -36,135 +48,205 @@
V_CVT_F64_I32
V_CVT_F64_U32
V_CVT_FLR_I32_F32
+V_CVT_I16_F16
V_CVT_I32_F32
V_CVT_I32_F64
+V_CVT_NORM_I16_F16
V_MAC_F32
-V_MAC_LEGACY_F32
-V_MADAK_F32
-V_MADI64_I32
-V_MADMK_F32
-V_MADU64_U32
-V_MAD_F32
+V_MAD_I16
+V_MAD_I32_I16
V_MAD_I32_I24
-V_MAD_LEGACY_F32
+V_MAD_I64_I32
+V_MAD_U16
+V_MAD_U32_U16
V_MAD_U32_U24
+V_MAD_U64_U32
+V_MAX3_F16
V_MAX3_F32
+V_MAX3_I16
V_MAX3_I32
+V_MAX3_U16
V_MAX3_U32
+V_MAX_F16
V_MAX_F32
V_MAX_F64
+V_MAX_I16
V_MAX_I32
-V_MAX_LEGACY_F32
+V_MAX_U16
V_MAX_U32
V_MBCNT_HI_U32_B32
V_MBCNT_LO_U32_B32
+V_MED3_F16
V_MED3_F32
V_MED3_I32
V_MED3_U32
+V_MIN3_F16
V_MIN3_F32
+V_MIN3_I16
V_MIN3_I32
+V_MIN3_U16
V_MIN3_U32
+V_MIN_F16
V_MIN_F32
V_MIN_F64
+V_MIN_I16
V_MIN_I32
-V_MIN_LEGACY_F32
+V_MIN_U16
V_MIN_U32
V_MOVRELD_B32
+V_MOVRELSD_2_B32
V_MOVRELSD_B32
V_MOVRELS_B32
V_MOV_B32
+V_MOV_FED_B32
V_MQSAD_PK_U16_U8
AMD GCNͱAMD RDNA2ͷ
ϕΫλԋࢉ໋ྩͷdiff
݁ߏͳ਺ͷ໋ྩ͕
৽͍͠RDNA2Ͱ͸
࡟আ͞Ε͍ͯΔ
GPU͸ಉ͡ϕϯμͰ͋ͬͯ΋
໋ྩηοτͷޓ׵ੑ͸ͳ͘ͳΓ͕ͪ

View Slide

GPU Aͷ
࣮ߦՄೳόΠφϦ
GPU A GPU B GPU C
GPUͷ࣮ߦՄೳόΠφϦΛ
௚઀༻ҙ࣮ͯ͠ߦ͢Δͱ
ಛఆͷGPUͰ͔͠ಈ͔ͳ͘ͳΔ
ϋʔυ΢ΣΞΛݶఆͰ͖ΔՈఉ༻ήʔϜػ͸͜ΕΛ΍͍ͬͯΔ
࣮ߦ࣌
ίϯύΠϧ࣌

View Slide

void main() {
vec3 normal =
normalize( inpu
t_normal.xyz );
vec3 pos =
input_position.
xyz;
vec3 N =
normal;
GPU A GPU B GPU C
GLSL(ߴڃݴޠ)
࣮ߦ࣌
ίϯύΠϧ࣌
OpenGLͷ৔߹
࣮ߦ࣌ʹγΣʔμΛ
ίϯύΠϧ͢Δ
͕͔͔࣌ؒΔ

View Slide

void main() {
vec3 normal =
normalize( inpu
t_normal.xyz );
vec3 pos =
input_position.
xyz;
vec3 N =
normal;
ߴڃݴޠ
a b
×
+
3
a b
×
+
3
࣮ߦՄೳόΠφϦ
AST
AST
ࣈ۟ղੳ
ߏจղੳ
λʔήοτ
ඇґଘͷ
࠷దԽ
λʔήοτ
όΠφϦͷ
ੜ੒
ίϯύΠϥͷॲཧ͸େ͖͘෼͚ͯ4ஈ֊
a b
×
+
3
AST
λʔήοτ
ݻ༗ͷ
࠷దԽ

View Slide

void main() {
vec3 normal =
normalize( inpu
t_normal.xyz );
vec3 pos =
input_position.
xyz;
vec3 N =
normal;
ߴڃݴޠ
a b
×
+
3
a b
×
+
3
࣮ߦՄೳόΠφϦ
AST
AST
ࣈ۟ղੳ
ߏจղੳ
λʔήοτ
ඇґଘͷ
࠷దԽ
λʔήοτ
όΠφϦͷ
ੜ੒
a b
×
+
3
AST
λʔήοτ
ݻ༗ͷ
࠷దԽ
͜ͷ෦෼͸GPUຖʹߦ͏ඞཁ͕͋ΔͷͰ
࣮ߦ࣌ʹ΍Β͟ΔΛಘͳ͍
͜ͷ෦෼͸
ࣄલʹย෇͚ͯ΋໰୊ͳ͍
a b
×
+
3
͜ͷஈ֊ͷASTΛ
όΠφϦܗࣜͰ
γϦΞϥΠζ͓ͯ࣋ͬͯ͜͠͏

View Slide

void main() {
vec3 normal =
normalize( inpu
t_normal.xyz );
vec3 pos =
input_position.
xyz;
vec3 N =
normal;
ߴڃݴޠ
a b
×
+
3
a b
×
+
3
࣮ߦՄೳόΠφϦ
AST
AST
ࣈ۟ղੳ
ߏจղੳ
λʔήοτ
ඇґଘͷ
࠷దԽ
λʔήοτ
όΠφϦͷ
ੜ੒
a b
×
+
3
AST
λʔήοτ
ݻ༗ͷ
࠷దԽ
͜ͷ෦෼͸
ࣄલʹย෇͚ͯ΋໰୊ͳ͍
a b
×
+
3
SPIR-V
͜ͷஈ֊ͷASTΛ
όΠφϦܗࣜͰ
γϦΞϥΠζ͓ͯ࣋ͬͯ͜͠͏

View Slide

void main() {
vec3 normal =
normalize( inpu
t_normal.xyz );
vec3 pos =
input_position.
xyz;
vec3 N =
normal;
GPU A GPU B GPU C
GLSL(ߴڃݴޠ)
࣮ߦ࣌
ίϯύΠϧ࣌
Vulkanͷ৔߹
a b
×
+
3
glslc
SPIR-V
vkCreateShaderModule

View Slide

#version 450
#extension GL_ARB_separate_shader_objects : enable
#extension GL_ARB_shading_language_420pack : enable
#extension GL_KHR_shader_subgroup_basic : enable
#extension GL_KHR_shader_subgroup_arithmetic : enable
layout(local_size_x_id = 1, local_size_y_id = 2 ) in;
layout(std430, binding = 1) buffer layout1 {
float output_data[];
};
layout(constant_id = 3) const float value = 1;
void main() {
const uint x = gl_GlobalInvocationID.x;
const uint y = gl_GlobalInvocationID.y;
const uint width = gl_WorkGroupSize.x * gl_NumWorkGroups.x;
const uint index = x + y * width;
output_data[ index ] += value;
}
؆୯ͳGLSLͷྫ

View Slide

#version 450
};
void main() {
}
όοϑΝ

View Slide

#version 450
};
void main() {
}
εϨουID͔Β
όοϑΝͷͲ͜ʹॻ͔ܾ͘ΊΔ

View Slide

#version 450
};
void main() {
}
όοϑΝͷ1ཁૉʹ1ΛՃ͑Δ
value͸1
࣮ߦ͢Δ౓ʹόοϑΝͷ஋ΛΠϯΫϦϝϯτ͢Δ

View Slide

#version 450
};
void main() {
}
binding = 1ͷόοϑΝΛ
output_dataͱ݁ͼ͚ͭΔ
binding = 1ͷόοϑΝͬͯͲͷόοϑΝͷ͜ͱ?

View Slide

σεΫϦϓληοτ
όοϑΝ# CJOEJOH
όοϑΝ" CJOEJOH
όοϑΝ$ CJOEJOH
⋮
όοϑΝA
όοϑΝB
όοϑΝC
#version 450
#extension GL_ARB_shading_language_420pack : enabl
#extension GL_KHR_shader_subgroup_arithmetic : ena
layout(local_size_x_id = 1, local_size_y_id = 2 )
};
void main() {
const uint width = gl_WorkGroupSize.x * gl_NumWo
}
ॻ͖ࠐΈ
γΣʔμͷbindingͱvkCreateBufferͰ࡞ͬͨόοϑΝΛରԠ෇͚Δ
vkUpdateDescriptorSetsͰొ࿥

View Slide

σεΫϦϓλϓʔϧ
σεΫϦϓληοτ
⋮
όοϑΝA
όοϑΝB
όοϑΝC
σεΫϦϓληοτ͸
ϋʔυ΢ΣΞͷ
ݶΒΕͨϨδελΛ
࢖͏Մೳੑ͕͋Δ
σεΫϦϓληοτ
⋮
⋯
σεΫϦϓληοτ͸σεΫϦϓλϓʔϧ͔ΒׂΓ౰ͯΔ
vkAllocateDescriptorSets
ཁΒͳ͘ͳͬͨΒ
vkFreeDescriptorSets
Ͱฦ٫

View Slide

σεΫϦϓλϓʔϧ
σεΫϦϓληοτ
όοϑΝA
όοϑΝB
όοϑΝC
σεΫϦϓληοτ
⋮
⋯
σεΫϦϓληοτϨΠΞ΢τ
όοϑΝ༻ͷσεΫϦϓλ͕3ݸ͋ΔΑ͏ͳ
σεΫϦϓληοτΛ͍ͩ͘͞
ԿΛରԠ͚ͮΔҝͷ
σεΫϦϓλ͕
Կݸ༻ҙ͞Ε͍ͯΔ
σεΫϦϓληοτ͕
ཉ͍͔͠Λද͢

View Slide

σεΫϦϓλϓʔϧ
σεΫϦϓληοτ
όοϑΝA
όοϑΝB
όοϑΝC
σεΫϦϓληοτ
⋮
⋯
όοϑΝ༻ͷσεΫϦϓλ͕3ݸ͋ΔΑ͏ͳ
σεΫϦϓληοτΛ͍ͩ͘͞
ԿΛରԠ͚ͮΔҝͷ
σεΫϦϓλ͕
Կݸ༻ҙ͞Ε͍ͯΔ
σεΫϦϓληοτ͕
ཉ͍͔͠Λද͢
SPIR-VΛ
ಡΜͩΒΘ͔ΔͷͰ͸
a b
×
+
3

View Slide

SPIR-VΛ
ಡΜͩΒΘ͔ΔͷͰ͸
a b
×
+
3
Q.
A. Θ͔Δ
ͳͷͰSPIR-V͔ΒbindingΛ
ړΔϥΠϒϥϦ͕͋Δ
SPIRV-Reflect
https://github.com/KhronosGroup/SPIRV-Reflect
ϕϯμʔຖͷGPUͷυϥΠόʹ
͜ͷػೳΛ࣮૷͠ͳͯ͘ྑ͍

View Slide

γΣʔμϞδϡʔϧͱσεΫϦϓληοτϨΠΞ΢τΛ͚ͬͭ͘Δ
ͬͭ͘͘=์ஔ͞ΕΔbinding͸ଘࡏ͠ͳ͍
ίϯϐϡʔτύΠϓϥΠϯ
VkResult vkCreateComputePipelines(
VkDevice device,
VkPipelineCache pipelineCache,
uint32_t createInfoCount,
const VkComputePipelineCreateInfo* pCreateInfos,
VkPipeline* pPipelines
); typedef struct VkComputePipelineCreateInfo {
const void* pNext;
VkPipelineCreateFlags flags;
VkPipelineShaderStageCreateInfo stage;
VkPipelineLayout layout;
VkPipeline basePipelineHandle;
int32_t basePipelineIndex;
} VkComputePipelineCreateInfo;

View Slide

typedef struct VkComputePipelineCreateInfo {
const void* pNext;
VkPipelineCreateFlags flags;
typedef struct VkPipelineShaderStageCreateInfo {
const void* pNext;
VkPipelineShaderStageCreateFlags flags;
VkShaderStageFlagBits stage;
VkShaderModule module;
const char* pName;
const VkSpecializationInfo* pSpecializationInfo;
} VkPipelineShaderStageCreateInfo;
γΣʔμ
Ϟδϡʔϧ

View Slide

VkResult vkCreatePipelineLayout(
VkDevice device,
const VkPipelineLayoutCreateInfo* pCreateInfo,
VkPipelineLayout* pPipelineLayout
); typedef struct VkPipelineLayoutCreateInfo {
const void* pNext;
VkPipelineLayoutCreateFlags flags;
uint32_t setLayoutCount;
const VkDescriptorSetLayout* pSetLayouts;
uint32_t pushConstantRangeCount;
const VkPushConstantRange* pPushConstantRanges;
} VkPipelineLayoutCreateInfo;
σεΫϦϓλ
ηοτ
ϨΠΞ΢τ

View Slide

ύΠϓϥΠϯΩϟογϡ
VkResult vkCreateComputePipelines(
VkDevice device,
);
Ұ౓࡞ͬͨ ࣮ߦՄೳόΠφϦ౳Λ͓֮͑ͯ͘
͜Ε
Ҏલͱಉ͡಺༰ͰύΠϓϥΠϯͷ࡞੒Λཁٻ͞ΕͨΒ
Ωϟογϡͷ಺༰Λ࢖͏

View Slide

);
VkResult vkCreatePipelineCache(
VkDevice device,
const VkPipelineCacheCreateInfo* pCreateInfo,
VkPipelineCache* pPipelineCache
); typedef struct VkPipelineCacheCreateInfo {
const void* pNext;
VkPipelineCacheCreateFlags flags;
size_t initialDataSize;
const void* pInitialData;
} VkPipelineCacheCreateInfo;

View Slide

VkResult vkCreatePipelineCache(
VkDevice device,
const VkPipelineCacheCreateInfo* pCreateInfo,
VkPipelineCache* pPipelineCache
); typedef struct VkPipelineCacheCreateInfo {
const void* pNext;
VkPipelineCacheCreateFlags flags;
size_t initialDataSize;
const void* pInitialData;
} VkPipelineCacheCreateInfo;
VkResult vkGetPipelineCacheData(
VkDevice device,
size_t* pDataSize,
void* pData
);
ೋ࣍هԱ
࣍ճىಈ࣌͸
γΣʔμͷ
ίϯύΠϧΛճආ

View Slide

[v1
, v2
, v3
, v4
, v5
, v6
, v7
, v8
, v9
, v10]
͋ͱඞཁͳͷ͸ԿεϨουͰ࣮ߦ͢Δ͔
void vkCmdDispatch(
VkCommandBuffer commandBuffer,
uint32_t groupCountX,
uint32_t groupCountY,
uint32_t groupCountZ
);
͜ͷίϚϯυόοϑΝʹ
ݸͷεϨουͰ࣮ߦΛ։࢝͢ΔཁٻΛੵΉ
groupCountx
× groupCounty
× groupCountz
͜ͷίϚϯυΛΩϡʔʹྲྀ͢ͱGPUͰγΣʔμ͕࣮ߦ͞ΕΔ

View Slide

ίϚϯυόοϑΝ
vkCmdDispatch
vkCmdDispatch
vkCmdDispatchΛ
ෳ਺Ωϡʔʹྲྀͨ͠৔߹
ͦΕΒ͕
ॱ൪ʹ࣮ߦ͞ΕΔอূ͸ͳ͍
GPUͷϓϩηοαʹ༨༟͕͋Δ৔߹
ෳ਺ͷvkCmdDispatch͕
ಉ࣌ʹ࣮ߦ͞ΕΔ͜ͱ΋͋Δ
stallͨ͠vkCmdDispatch͕
ޙճ͠ʹͳΔ͜ͱ΋͋Δ
32εϨου
64εϨου

View Slide

ίϚϯυόοϑΝ ෳ਺ͷvkCmdDispatchͷؒʹ
σʔλͷґଘؔ܎͕͋Δ৔߹͸
vkCmdPipelineBarrierͰ
ґଘؔ܎Λ໌ࣔ͢Δͱ
ద੾ͳॱংͰ࣮ߦ͞ΕΔ
vkCmdPipelineBarrier
vkCmdDispatch
vkCmdDispatch

View Slide

void vkCmdPipelineBarrier(
VkPipelineStageFlags srcStageMask,
VkPipelineStageFlags dstStageMask,
VkDependencyFlags dependencyFlags,
uint32_t memoryBarrierCount,
const VkMemoryBarrier* pMemoryBarriers,
uint32_t bufferMemoryBarrierCount,
const VkBufferMemoryBarrier* pBufferMemoryBarriers,
uint32_t imageMemoryBarrierCount,
const VkImageMemoryBarrier* pImageMemoryBarriers
);
typedef struct VkBufferMemoryBarrier {
const void* pNext;
VkAccessFlags srcAccessMask;
VkAccessFlags dstAccessMask;
uint32_t srcQueueFamilyIndex;
uint32_t dstQueueFamilyIndex;
VkBuffer buffer;
VkDeviceSize offset;
VkDeviceSize size;
} VkBufferMemoryBarrier;
͜ͷόοϑΝ

View Slide

);
typedef struct VkBufferMemoryBarrier {
const void* pNext;
VkBuffer buffer;
VkDeviceSize offset;
VkDeviceSize size;
} VkBufferMemoryBarrier;
͜ͷόοϑΝ
όϦΞͷલʹ͜ͷόοϑΝΛ৮ͬͨίϚϯυ͕׬ྃ͢Δ·Ͱ
όϦΞͷޙͰ͜ͷόοϑΝΛ৮ΔίϚϯυΛ։࢝ͯ͠͸͍͚·ͤΜ

View Slide

{
auto mapped = staging_buffer->map< float >();
std::fill( mapped.begin(), mapped.end(), 0.f );
}
{
auto rec = command_buffer->begin();
rec.copy( staging_buffer, device_local_buffer );
rec.barrier(
vk::AccessFlagBits::eTransferWrite,
vk::AccessFlagBits::eShaderRead,
vk::PipelineStageFlagBits::eTransfer,
vk::PipelineStageFlagBits::eComputeShader,
vk::DependencyFlagBits( 0 ),
{ device_local_buffer },
{}
);
rec.bind_descriptor_set(
vk::PipelineBindPoint::eCompute,
pipeline_layout,
descriptor_set
);
θϩΫϦΞͨ͠
ϝϞϦΛ
GPUʹૹͬͯ
ίϐʔ׬ྃΛ
଴͔ͬͯΒ

View Slide

rec.bind_descriptor_set(
pipeline_layout,
descriptor_set
);
rec.bind_pipeline(
pipeline
);
rec->dispatch( 4, 2, 1 );
rec.barrier(
vk::AccessFlagBits::eShaderWrite,
vk::AccessFlagBits::eTransferRead,
{}
);
rec.copy( device_local_buffer, staging_buffer );
}
σεΫϦϓληοτΛ
ࢦఆͯ͠
ύΠϓϥΠϯΛ
ࢦఆͯ͠
࣮ߦͯ͠
࣮ߦͷ׬ྃΛ
଴͔ͬͯΒ

View Slide

{}
);
rec.copy( device_local_buffer, staging_buffer );
}
command_buffer->execute(
gct::submit_info_t()
);
command_buffer->wait_for_executed();
std::vector< float > host;
host.reserve( 1024 );
{
auto mapped = staging_buffer->map< float >();
std::copy( mapped.begin(), mapped.end(), std::back_inserter( host ) );
}
unsigned int count;
nlohmann::json json = host;
std::cout << json.dump( 2 ) << std::endl;
CPUଆʹίϐʔ
JSONʹͯ͠μϯϓ
͜͜·Ͱͷ಺༰ΛΩϡʔʹྲྀͯ͠
ίϚϯυͷ׬ྃΛ଴ͬͯ
GPU͔Βདྷͨ
σʔλΛ

View Slide

$ ./src/compute
[
1.0,
1.0,
1.0,
1.0,
1.0,
1.0,
1.0,
1.0,
1.0,
...
1.0,
1.0,
1.0,
1.0,
1.0,
1.0,
1.0
]
શ෦ΠϯΫϦϝϯτ͞ΕͯΔ

View Slide

Graphics
Processing
Unit
Α͘๨ΕΒΕΔ͕
GPUͷG͸
GraphicsͷG

View Slide

vkBindBufferMemory
vkBindImageMemory
VkDeviceMemory VkImage
͜ͷϝϞϦͷத਎͸൚༻తͳܭࢉσʔλͰ͢
͜ͷϝϞϦͷத਎͸ը૾Ͱ͢
VkImageͰϝϞϦʹஔ͔Εͨσʔλ͕
ը૾Ͱ͋Δͱ͍͏͜ͱΛ໌ࣔ͢Δ

View Slide

vkBindBufferMemory
vkBindImageMemory
VkDeviceMemory VkImage
σʔλ͸CPU͔ΒૹΒΕͨ··ͷॱংͰ
GPUʹஔ͔Ε·͢
σʔλ͸ը૾ͷ༻్ʹԠͯ͡࠷దͳஔ͖ํʹ
ม׵ͯ͠GPUʹஔ͔Ε·͢
VkImageʹը૾ͷ༻్Λࢦఆ͢Δͱ
Vulkan͸ͦͷ༻్ʹదͨ͠ฒͼํͰϝϞϦʹϐΫηϧΛฒ΂Δ

View Slide

p
ྫ͑͹ΠϝʔδΛςΫενϟͱͯ͠࢖͏৔߹
p
ͷҐஔͷ৭Λܾఆ͢Δͷʹ
࠷ۙ๣ͳΒ ͷϐΫηϧΛ
ઢܗิ׬ͳΒ ͱ ͷϐΫηϧΛ
Cubicิ׬ͳΒ ͱ ͷϐΫηϧΛ
ͱ
ಡΉඞཁ͕͋Δ

View Slide

ΠϝʔδΛx࣠ํ޲ʹ1ߦͮͭ
ϝϞϦʹஔ͍͍ͯΔͱ
͜ͷൣғͷ஋͕ඞཁ
y࣠ํ޲ͷྡ઀͢ΔϐΫηϧ͕
ϝϞϦ্Ͱ཭ΕͨҐஔʹه࿥͞ΕΔ
࣍ʹಡΉϐΫηϧ͕
Ωϟογϡʹ৐͍ͬͯΔ֬཰͕Լ͕Δ

View Slide

ྫ͑͹ΠϝʔδͷϐΫηϧ͕
͜Μͳॱ൪ͰϝϞϦʹฒΜͰ͍Δͱ
͋ΔϐΫηϧͷ஋ΛಡΜͩޙͰ
ۙ๣ͷϐΫηϧΛಡΜͩ࣌ʹ
ͦͷϐΫηϧ͕
Ωϟογϡʹ৐͍ͬͯΔ֬཰্͕͕Δ

View Slide

VkResult vkCreateImage(
VkDevice device,
const VkImageCreateInfo* pCreateInfo,
VkImage* pImage
);
typedef struct VkImageCreateInfo {
const void* pNext;
VkImageCreateFlags flags;
VkImageType imageType;
VkFormat format;
VkExtent3D extent;
uint32_t mipLevels;
uint32_t arrayLayers;
VkSampleCountFlagBits samples;
VkImageTiling tiling;
VkImageUsageFlags usage;
VkImageLayout initialLayout;
} VkImageCreateInfo;
༻్
VkImage࡞੒࣌ʹ
༻్Λࢦఆ͢Δ
༻్͸ϏοτϑϥάͰ
ෳ਺ࢦఆͯ͠΋ྑ͍
VK_IMAGE_USAGE_TRANSFER_DST_BIT|
VK_IMAGE_USAGE_SAMPLED_BIT
ྫ
vkCopyImageͷड͚ଆ͔ͭ
ςΫενϟαϯϓϦϯάର৅

View Slide

void vkCmdPipelineBarrier(
VkPipelineStageFlags srcStageMask,
VkPipelineStageFlags dstStageMask,
);
typedef struct VkImageMemoryBarrier {
const void* pNext;
VkImageLayout oldLayout;
VkImageLayout newLayout;
VkImage image;
VkImageSubresourceRange subresourceRange;
} VkImageMemoryBarrier;
͜ͷΠϝʔδΛ
͜ͷϨΠΞ΢τ͔Β
͜ͷϨΠΞ΢τʹ
όϦΞ͢Δ͍ͭͰʹ
ΠϝʔδͷϨΠΞ΢τΛ
มߋͰ͖Δ

View Slide

ίϚϯυόοϑΝ
ը૾Λੜ੒
CPUଆʹίϐʔ
VK_IMAGE_LAYOUT_GENERALͰు͘
VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMALͰཉ͍͠
VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMALʹม׵
GPU͕ಡΈॻ͖͢Δͷʹదͨ͠ϨΠΞ΢τ
సૹ͢Δͷʹదͨ͠ϨΠΞ΢τ
సૹ͢Δͷʹదͨ͠ϨΠΞ΢τ
λΠϧແޮ͔ͭ
ϨΠϠʔ͕1ຕ͔ͭ
mipmapͳ͔ͭ͠
సૹʹదͨ͠ϨΠΞ΢τ
= ߦϝδϟʔͰ
ύσΟϯάͤͣʹ
ॱ൪ʹϐΫηϧ͕ฒΜͩ
ϨΠΞ΢τ
CPU͔ΒಡΈ΍͍͢

View Slide

#version 450
};
layout(set = 0, binding = 0, rgba8) uniform writeonly image2D img;
void main() {
...
imageStore( img, ivec2( pos.xy ), color );
}
Storage ImageΛ࢖͏ͱ
ίϯϐϡʔτύΠϓϥΠϯ͔ΒΠϝʔδΛಡΈॻ͖Ͱ͖Δ
color͸pos.xyͷҐஔͷϐΫηϧ͕ஔ͔ΕΔ΂͖Ґஔʹॻ͔ΕΔ

View Slide

PolyMorph
େ͖ͳࡾ֯ܗΛ
খ͞ͳෳ਺ͷࡾ֯ܗʹ
෼ׂ͢Δ
(ςοηʔϨʔλ)
GPUʹ͸
ޮ཰Α͘3DάϥϑΟΫεΛඳ͘ҝͷ
ઐ༻ͷϋʔυ΢ΣΞ͕৭ʑࡌ͍ͬͯΔ

View Slide

ϥελϥΠβ
3ͭͷ௖఺Ͱఆٛ͞Εͨࡾ֯ܗ͕
ͲͷϐΫηϧʹରԠ͢Δ͔ΛٻΊΔ
GPUʹ͸

View Slide

Raster Operators
γΣʔσΟϯάͷ݁ՌΛू໿ͯ͠
࠷ऴతͳΠϝʔδʹه࿥͢Δ৭Λܾఆ͢Δ
GPUʹ͸

View Slide

GPU
+
+
21ੈلͷ
େྔͷ +
ϓϩηοαͰ͸
ޮ཰͕ѱ͍෦෼Λ
ิ͏ϋʔυ΢ΣΞ

View Slide

Input Assembly
Vertex Shader
Tessellation Control Shader
Tessellation
Tessellation Evaluation Shader
Geometry Shader
Rasterization
Fragment Shader
Color Blend
ϋʔυ΢ΣΞ
ϋʔυ΢ΣΞ
ϋʔυ΢ΣΞ
3DάϥϑΟΫεͷ
ඳըखॱͷॴʑͰ
ઐ༻ͷϋʔυ΢ΣΞΛ
࢖͍͍ͨ
ϋʔυ΢ΣΞ

View Slide

Input Assembly
Vertex Shader
Tessellation
Geometry Shader
Rasterization
Fragment Shader
Color Blend
ϋʔυ΢ΣΞ
ϋʔυ΢ΣΞ
ϋʔυ΢ΣΞ
࢒ΓͷεςοϓͦΕͧΕʹ
SPIR-VΛ݁ͼ͚ͭΔ
a b
×
+
3
a b
×
+
3
a b
×
+
3
a b
×
+
3
a b
×
+
3
a b
×
+
3
ϋʔυ΢ΣΞ

View Slide

άϥϑΟΫε
ύΠϓϥΠϯ

View Slide

Input Assembly
Vertex Shader
Tessellation
Geometry Shader
Rasterization
Fragment Shader
Color Blend

View Slide

Input Assembly
Vertex Shader
Tessellation
Geometry Shader
Rasterization
Fragment Shader
Color Blend
࣮ߦ࣌ʹಈతʹมߋͰ͖Δ
ඞཁ͕͋ΔઃఆΛࢦఆ͢Δ

View Slide

Ϩϯμʔύε
ͱ͸

View Slide

Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Ϩϯμʔύε
ෳ਺ͷάϥϑΟΫεύΠϓϥΠϯΛଋͶͨ΋ͷ

View Slide

Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
ϚϧνύεϨϯμϦϯά
VkImage
1ஈ֊໨ͷϨϯμϦϯάͷ݁ՌΛ
ೖྗͱͯ͠2ஈ֊໨ͷϨϯμϦϯάΛߦ͏
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend

View Slide

Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
VkImage
VkImage
࠲ඪ
๏ઢ
ਂ౓
VkImage
ࡐ࣭
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
Input Assemb
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
র໌ র໌ র໌
GόοϑΝ

View Slide

VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
র໌ র໌ র໌
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
∑
VkImage
ϨϯμϦϯά݁Ռ

View Slide

VS
TCS
sellation
TES
GS
erization
FS
or Blend
Image
Image
Image
Image
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
র໌ র໌ র໌
In
R
∑
V
ϨϯμϦϯά݁Ռ
͜͜Ͱશͯͷর໌Λ
ॱʹܭࢉ͢ΔΑΓεέʔϧ͢Δ

View Slide

Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
VkImage
VkImage
࠲ඪ
๏ઢ
ਂ౓
VkImage
ࡐ࣭
VkImage
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkIma
VS
TCS
Tessellati
TES
GS
Rasterizat
FS
Color Ble
ϨϯμϦ
GόοϑΝʹ࢒Βͳ͔ͬͨ(=ଞͷ΋ͷͷഎޙʹ͋ͬͯݟ͑ͳ͍)
ϐΫηϧ͸ҎޙͷܭࢉʹݱΕͳ͍

View Slide

Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
VkImage
VkImage
࠲ඪ
๏ઢ
ਂ౓
VkImage
ࡐ࣭
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkIm
Input A
V
TC
Tesse
TE
G
Raste
F
Color
র໌ র໌
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
র໌1ͷҐஔ͔Β
ϨϯμϦϯά
VkImage
ਂ౓
র໌1ͷҐஔ͔Βͷ
ϨϯμϦϯά݁Ռʹө͍ͬͯͳ͍ͳΒ
ͦ͜ʹ͸র໌1ͷޫ͕ಧ͔ͳ͍

View Slide

VkImage
TES
GS
Rasterization
FS
Color Blend
VkImage
TES
GS
Rasterization
FS
Color Blend
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
ϨϯμϦϯά݁Ռʹը૾ॲཧΛߦ͏
ϨϯμϦϯά݁Ռ
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkImage
ඃࣸքਂ౓ޮՌ
τʔϯϚοϓͳͲ
ը૾ॲཧ͞ΕͨϨϯμϦϯά݁Ռ

View Slide

ίϚϯυόοϑΝ
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
όϦΞͰ
ෳ਺ͷάϥϑΟΫεύΠϓϥΠϯͷ࣮ߦʹ
ґଘؔ܎Λ࣋ͨͤΕ͹ྑ͍ͷͰ͸
͜ͷํ๏Ͱ΋Ͱ͖Δ
͔͜͠͠ͷํ๏Ͱ͸
ϞόΠϧGPUͰੑೳ͕ग़ͳ͍
ύΠϓϥΠϯΛ࣮ߦ
ύΠϓϥΠϯΛ࣮ߦ

View Slide

CPU GPU
ࡉ͍
ϞόΠϧGPU

View Slide

CPU GPU
ࡉ͍
ଠ͍
1ը໘෼ͷ
ϨϯμϦϯά݁ՌΛஔ͘ʹ͸
খ͗͢͞Δ
VkImage
ϨϯμϦϯά݁Ռ͸
͜͜ʹஔ͔͘͠ͳ͍
SRAM

View Slide

CPU GPU
ࡉ͍
ଠ͍
ը໘ͷҰ෦͚ͩΛ
SRAM্ͰϨϯμϦϯά͢Δ
SRAM
ॱ൪ʹϨϯμϦϯάͯ݁͠ՌΛॻ͖ࠐΉ
λΠϧ

View Slide

CPU GPU
ࡉ͍
ଠ͍
SRAM
1 1 2
όϦΞ
1ύε໨Λ1ը໘෼ϝΠϯϝϞϦʹు͍͔ͯΒ
ϝΠϯϝϞϦΛಡΜͰ2ύε໨Λܭࢉ࢝͠ΊΔ
όϦΞΛ࢖ͬͨ
Ϛϧνύεͷ৔߹

View Slide

Ϩϯμʔύε
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Ϩϯμʔύε಺ͷෳ਺ͷύΠϓϥΠϯ͸
ೖग़ྗʹґଘؔ܎Λ࣋ͨͤΔ͜ͱ͕Ͱ͖Δ
ͨͩ͠B΍Cͷ ͷϐΫηϧΛܭࢉ͢Δ࣌
ಡΊΔ͜ͱ͕อূ͞ΕΔͷ͸Aͷ ͷҐஔͷ஋͚ͩ
(x, y)
(x, y)
"
# $

View Slide

CPU GPU
ࡉ͍
ଠ͍
SRAM
1 2
Ϩϯμʔύεͷ
৔߹
1ͭͷλΠϧʹର͢Δ
ෳ਺ͷύΠϓϥΠϯͷॲཧΛ
Ұ౓ʹ࣮ߦ
ϝΠϯϝϞϦ΁ͷ
ॻ͖ࠐΈ͸
࠷ޙͷ1౓͚ͩ

View Slide

ό
Ϧ
Ξ
ό
Ϧ
Ξ

View Slide

ίϚϯυόοϑΝ
ύΠϓϥΠϯ୯ҐͰ͸ͳ͘
Ϩϯμʔύε୯ҐͰ࣮ߦ͢Δ
Ϩϯμʔύε1Λ࣮ߦ
όϦΞ
όϦΞ

View Slide

GPU
+
+
21ੈلͷ
ϨϯμϦϯά݁ՌΛը໘ʹग़͍ͨ͠

View Slide

͜͜ʹॻ͘ͱग़Δ ίϯϙδλ
X Window System
Wayland Compositor
Windows DWM
etc.
Vulkan
ΞϓϦέʔγϣϯ
ը໘ʹૹΔө૾Λॻ͖ࠐΉҝͷϝϞϦ͸
ଟ͘ͷ৔߹ίϯϙδλ͕઎༗͍ͯ͠Δ

View Slide

X Window System
Wayland Compositor
Windows DWM
etc.
Vulkan
ΞϓϦέʔγϣϯ
ΞϓϦέʔγϣϯ͸ίϯϙδλ͔Β
ඳը಺༰Λ౉͢ઌαʔϑΣεΛ໯͏
ඳը಺༰ͷॻ͖ࠐΈઌ͍ͩ͘͞
͜͜ʹඳը಺༰Λ
౉͍ͯͩ͘͠͞
αʔϑΣε

View Slide

ΞϓϦέʔγϣϯ͸ίϯϙδλ͔Β
ඳը಺༰Λ౉͢ઌαʔϑΣεΛ໯͏
ϓϥοτϑΥʔϜݻ༗ͷϋϯυϥͰ
Windows
X11
Wayland
Android
Fuchsia
iOS
GGP
Nintendo Switch
HWND
xcb_window_t*
wl_surface*
ANativeWindow*
zx_handle_t
CAMetalLayer*
GgpStreamDescriptor
void*

View Slide

HWND
xcb_window_t*
wl_surface*
ANativeWindow*
zx_handle_t
CAMetalLayer*
GgpStreamDescriptor
void*
vkCreateWin32SurfaceKHR
vkCreateImagePipeSurfaceFUCHSIA
VkSurfaceKHR
vkGetPhysicalDeviceXcbPresentationSupportKHR
vkCreateIOSSurfaceMVK
vkGetPhysicalDeviceWaylandPresentationSupportKHR
vkCreateStreamDescriptorSurfaceGGP
vkGetPhysicalDeviceWaylandPresentationSupportKHR
vkCreateViSurfaceNN

View Slide

͜͜ʹॻ͘ͱग़Δ
Vulkan
ΞϓϦέʔγϣϯ
ॻ͍ͯΔ
ಡΜͰΔ
ίϯϙδλ
ॻ͍ͯΔ
ίϯϙδλ͕ಡΜͰ͍ΔϝϞϦʹ௚઀ॻ͘ͱ
ඳ͍͍ͯΔ్தͷ΋ͷ͕ը໘ʹग़ͯ͠·͏

View Slide

͜͜ʹॻ͘ͱग़Δ
Vulkan
ΞϓϦέʔγϣϯ
ॻ͍ͯΔ
ಡΜͰΔ
ίϯϙδλ
ॻ͍ͯΔ
ॻ͚ͨΒ
੾Γସ͑
੾ΓସΘͬͨΒ
ݹ͍ͷΛճऩ
εϫοϓ
νΣʔϯ

View Slide

VkResult vkCreateSwapchainKHR(
VkDevice device,
const VkSwapchainCreateInfoKHR* pCreateInfo,
VkSwapchainKHR* pSwapchain
);
typedef struct VkSwapchainCreateInfoKHR {
const void* pNext;
VkSwapchainCreateFlagsKHR flags;
VkSurfaceKHR surface;
uint32_t minImageCount;
VkFormat imageFormat;
VkColorSpaceKHR imageColorSpace;
VkExtent2D imageExtent;
uint32_t imageArrayLayers;
VkImageUsageFlags imageUsage;
VkSharingMode imageSharingMode;
VkSurfaceTransformFlagBitsKHR preTransform;
VkCompositeAlphaFlagBitsKHR compositeAlpha;
VkPresentModeKHR presentMode;
VkBool32 clipped;
VkSwapchainKHR oldSwapchain;
} VkSwapchainCreateInfoKHR;
͜ͷຕ਺͘Ε
͜ͷαʔϑΣεʹ
౉ͨ͢Ίͷ
ΠϝʔδΛ

View Slide

εϫοϓνΣʔϯ
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
͜ͷΠϝʔδ͸
͜ͷϨΠΞ΢τʹ͔͠ͳΕ·ͤΜ
͜ͷϝϞϦ͸ίϯϙδλͷ
ϓϩηεͱڞ༗͞Ε͍ͯ·͢
εϫοϓνΣʔϯ͸
ϝϞϦׂ͕Γ౰ͯΒΕͨ
Πϝʔδͷଋ
ίϯϙδλͷ౎߹Ͱ
ϨΠΞ΢τ͕
ݶఆ͞Ε͍ͯΔ

View Slide

εϫοϓνΣʔϯ
VkImage
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
εϫοϓνΣʔϯͷ
Πϝʔδʹ޲͔ͬͯ
άϥϑΟΫεύΠϓϥΠϯͰ
ϨϯμϦϯά

View Slide

ϑϨʔϜόοϑΝ
νΣʔϯ
ge
ge
age
mage
VkDeviceMemory
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
άϥϑΟΫεύΠϓϥΠϯ͸
৭ͱਂ౓ͱεςϯγϧΛు͘
VkDeviceMemory
VkImage
ਂ౓ͱεςϯγϧΛड͚ΔΠϝʔδΛ
ࣗ෼Ͱ༻ҙͯ͠
εϫοϓνΣʔϯͷΠϝʔδͱ͚ͬͭͯ͘
ϑϨʔϜόοϑΝʹ͢Δ

View Slide

ϑϨʔϜόοϑΝ
VkDeviceMemory
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkDeviceMemory
VkImage
VkResult vkCreateFramebuffer(
VkDevice device,
const VkFramebufferCreateInfo* pCreateInfo,
VkFramebuffer* pFramebuffer
);
typedef struct VkFramebufferCreateInfo {
const void* pNext;
VkFramebufferCreateFlags flags;
VkRenderPass renderPass;
uint32_t attachmentCount;
const VkImageView* pAttachments;
uint32_t width;
uint32_t height;
uint32_t layers;
} VkFramebufferCreateInfo;
࢖͏Πϝʔδͷ
Ϗϡʔͷ഑ྻ

View Slide

ry
ry
VkDeviceMemory
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkResult vkQueuePresentKHR(
VkQueue queue,
const VkPresentInfoKHR* pPresentInfo
);
typedef struct VkPresentInfoKHR {
const void* pNext;
uint32_t swapchainCount;
const VkSwapchainKHR* pSwapchains;
const uint32_t* pImageIndices;
VkResult* pResults;
} VkPresentInfoKHR;
͜ͷεϫοϓνΣʔϯͷ
͜ͷΠϝʔδΛ
ίϯϙδλʹૹΕ
ඳ͚ͨΒ

View Slide

εϫοϓνΣʔϯ
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkResult vkAcquireNextImageKHR(
VkDevice device,
VkSwapchainKHR swapchain,
uint64_t timeout,
VkSemaphore semaphore,
VkFence fence,
uint32_t* pImageIndex
);
εϫοϓνΣʔϯͷΠϝʔδ΁ͷॻ͖ࠐΈ͸
ίϯϙδλଆ͕ย෇͍͔ͯΒߦ͏ඞཁ͕͋Δ
΋͏ॻ͚Δ?

View Slide

VkResult vkAcquireNextImageKHR(
VkDevice device,
VkSwapchainKHR swapchain,
uint64_t timeout,
VkSemaphore semaphore,
VkFence fence,
uint32_t* pImageIndex
);
VkResult vkCreateSemaphore(
VkDevice device,
const VkSemaphoreCreateInfo* pCreateInfo,
VkSemaphore* pSemaphore
);
typedef struct VkSubmitInfo {
const void* pNext;
const VkPipelineStageFlags* pWaitDstStageMask;
uint32_t commandBufferCount;
const VkCommandBuffer* pCommandBuffers;
uint32_t signalSemaphoreCount;
const VkSemaphore* pSignalSemaphores;
} VkSubmitInfo;
Πϝʔδͷ४උ͕Ͱ͖ͨΒ
͜ͷηϚϑΥʹ௨஌
ࠓ͔Βྲྀ͢ίϚϯυ͸
ηϚϑΥ΁ͷ௨஌Λ଴͔ͬͯΒ
࣮ߦͤΑ
Ωϡʔͷ֎΍Ωϡʔؒͷಉظ͸
όϦΞͰ͸ͳ͘ηϚϑΥΛ࢖͏

View Slide

View Slide

Vulkan
Modern Vulkan
NAOMASA MATSUBAYASHI
Twitter: @fadis_
͍·Ͳ͖ͷ

View Slide

Vulkan 1.1

View Slide

όοϑΝ" CJOEJOH
όοϑΝA
#version 450
#extension GL_EXT_shader_16bit_storage : require
uint16_t output_data[];
};
...
std::vector< std::uint16_t > data;
16bit੔਺ΛόοϑΝʹॻ͍ͯ
γΣʔμ͔Β16bit੔਺ͱͯ͠
ಡΉ
ܭࢉ͸32bit੔਺Ͱߦ͏
copy
16bitετϨʔδ

View Slide

typedef struct VkPhysicalDevice16BitStorageFeatures {
void* pNext;
VkBool32 storageBuffer16BitAccess;
VkBool32 uniformAndStorageBuffer16BitAccess;
VkBool32 storagePushConstant16;
VkBool32 storageInputOutput16;
} VkPhysicalDevice16BitStorageFeatures;
GPU͸16bitͷload/store͕Ͱ͖ͳ͍͔΋͠Εͳ͍
৽͘͠௥Ճ͞Εͨ
VkPhysicalDevice16BitStorageFeatures
Λௐ΂Ε͹
GPU͕ͦΕͧΕͷঢ়گͰ16bitͷload/storeΛͰ͖Δ͔͕Θ͔Δ
16bitετϨʔδ

View Slide

#version 450
float16_t output_data[];
};
...
16bitͷload/storeʹରԠ͍ͯ͠Δ৔߹
൒ਫ਼౓ුಈখ਺఺਺ͷload/store΋Ͱ͖Δ
#version 450
f16vec4 output_data[];
};
...
ϕΫλܕ΋OK
16bitετϨʔδ

View Slide

GPUͷϓϩηοα͸
32͔Β64ݸͷ஋ΛҰ౓ʹॲཧ͢Δ
SIMD໋ྩΛඋ͍͑ͯΔ
Vulkan͸͜ΕΛ32εϨουͱΧ΢ϯτ͠
1ݸͷ஋Λૢ࡞͢Δؔ਺32εϨουΛ
1ͭͷSIMD໋ྩͷ࣮ߦʹׂΓ౰ͯΔ
͜ͷ32εϨουΛSubgroupͱݺͿ
Subgroup Operation

View Slide

⋯
⋯ ⋯
+
+
+
+
+
ਨ௚Ճࢉ
ී௨ʹa+bΛ͢Δͱ
͜ΕʹͳΔ
a
b
Subgroup Operation

View Slide

⋯
⋯
⋯
⋯
ਫฏՃࢉ
+
+
+
+
a
subgroupAdd(a)
∑
n
an
Subgroup Operation

View Slide

⋯
⋯
⋯
⋯
ਫฏՃࢉ
+
+
+
+
a
subgroupInclusiveAdd(a)
Subgroup Operation

View Slide

⋯
⋯

⋯
⋯
ਫฏՃࢉ
+
+
+
a
subgroupExclusiveAdd(a)
+
Subgroup Operation

View Slide

⋯
⋯
⋯
ਫฏՃࢉ
+
a
subgroupClusteredAdd(a,2)
+ +
2ͭͮͭ
Subgroup Operation

View Slide

⋯

⋯ ⋯
γϟοϑϧ
subgroupShuffle(a,b)
a
b
͜ͷॱͰฒ΂ସ͑
Subgroup Operation

View Slide

⋯
⋯
⋯
ϒϩʔυΩϟετ
a
subgroupBroadcast(a,0)
શ෦ ʹͳΔ
a0
Subgroup Operation

View Slide

⋯
⋯
⋯
ϒϩʔυΩϟετ
a
subgroupQuadBroadcast(a)
4ͭͮͭ
Subgroup Operation

View Slide

struct VkPhysicalDeviceSubgroupProperties {
void* pNext;
uint32_t subgroupSize;
VkShaderStageFlags supportedStages;
VkSubgroupFeatureFlags supportedOperations;
VkBool32 quadOperationsInAllStages;
};
SubgroupͷαΠζΛҙࣝ͠ͳ͚Ε͹ͳΒͳ͘ͳͬͨ
औಘͰ͖ΔΑ͏ʹ͠Α͏
Subgroup Operation

View Slide

struct VkPhysicalDeviceSubgroupProperties {
void* pNext;
uint32_t subgroupSize;
VkShaderStageFlags supportedStages;
VkSubgroupFeatureFlags supportedOperations;
VkBool32 quadOperationsInAllStages;
};
GPUʹΑͬͯ͸શͯͷਫฏԋࢉΛαϙʔτͰ͖ͳ͍͔΋͠Εͳ͍
ͲΕ͕࢖͑Δ͔
ௐ΂ΒΕΔΑ͏ʹ
͠Α͏
Subgroup Operation

View Slide

͜ͷ
෺ཧσόΠε
+ Vulkan 1.0
VK_KHR_SWAPCHAIN_EXTENSION_NAME֦ு෇͖
= VkDevice
͜ͷόʔδϣϯͷ"1* ࿦ཧσόΠε

View Slide

͜Ε͸Vulkan 1.0Ͱ΋Ͱ͖Δ
ຕ໨ͷ
(16
+ Vulkan 1.0
= VkDevice
ຕ໨ͷ
(16
Vulkan 1.0
+ = VkDevice

View Slide

ຕ໨ͷ
(16
ຕ໨ͷ
(16
Vulkan 1.1 = VkDevice
%FWJDF(SPVQ
+
/7-JOL౳Ͱ઀ଓ͞Εͨෳ਺ͷ(16͔Β
ͭͷ࿦ཧσόΠεΛ࡞Δ
Device Group

View Slide

ຕ໨ͷ
(16
ຕ໨ͷ
(16
%FWJDF(SPVQ
ίϚϯυόοϑΝ
ίϚϯυ ίϚϯυ
Ωϡʔʹྲྀͨ͠ίϚϯυ͸%FWJDF(SPVQ಺ͷ
શͯͷ(16Ͱ࣮ߦ͞ΕΔ
Device Group

View Slide

ຕ໨ͷ
(16
ຕ໨ͷ
(16
%FWJDF(SPVQ
ίϚϯυόοϑΝ
ίϚϯυ ίϚϯυ
ίϚϯυόοϑΝ୯ҐͰ
࣮ߦ͢Δ(16Λ੍ݶͰ͖Δ
1ຕ໨ͷGPU͚ͩͰ࣮ߦ
Device Group

View Slide

ຕ໨ͷ
(16
ຕ໨ͷ
(16
%FWJDF(SPVQ
ίϚϯυόοϑΝ
ίϚϯυ
(16͸ෳ਺͚ͩͲ
Ωϡʔ͸ಉ͔ͩ͡Β
όϦΞͰಉظ͕Ͱ͖Δ
ίϚϯυόοϑΝ
ίϚϯυ
ίϚϯυόοϑΝ
όϦΞ
྆ํͰ࣮ߦ
Device Group

View Slide

VRͰ͸ϔουηοτͷϨϯζʹΑΔ࿪ΈΛ
ϨϯμϦϯάଆͰଧͪফ͢

View Slide

େ͖͘දࣔ͞ΕΔ=ղ૾౓͕ඞཁ
খ͘͞දࣔ͞ΕΔ=ղ૾౓Λ্͛ͯ΋ແବ

View Slide

୺ͷํ͚ͩ
࠷ॳ͔Βখ͘͞ඳ͜͏

View Slide

Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
Ϩϯμʔύε
ಉ͡௖఺഑ྻͷඳըཁٻΛ
Ϩϯμʔύεͷෳ਺ͷύΠϓϥΠϯʹҰ੪ʹྲྀ͢
ό
Ϧ
Ξ
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
มܗ
Multiview

View Slide

Unprotected Protected
1SPUFDUFEͳϝϞϦͷதͰ
࡞ΒΕͨσʔλ͸
(16ͷ֎ʹ࣋ͪग़ͤͳ͍
ίϐʔϓϩςΫτ͞Εͨը૾΍ಈը͕
(16ͷϝϞϦ͔ΒಡΈऔΒΕΔͷΛ
๷͍͗ͨͬΆ͍
Protected Memory

View Slide

Vulkan 1.2

View Slide

όοϑΝ" CJOEJOH
όοϑΝA
#version 450
uint8_t output_data[];
};
...
std::vector< std::uint8_t > data;
8bit੔਺ΛόοϑΝʹॻ͍ͯ
γΣʔμ͔Β8bit੔਺ͱͯ͠
ಡΉ
copy
8bitετϨʔδ
16bitಉ༷
8bit੔਺ͷϕΫλ
(ex. u8vec4)
΋OK

View Slide

8bitετϨʔδ
ͳΜͰ୹͍੔਺ͷαϙʔτΛ௥Ճ͢Δͷ
χϡʔϥϧωοτϫʔΫ͸
ݸʑͷॏΈͷਫ਼౓ΑΓ΋
ॏΈͷݸ਺͕
ੑೳʹେ͖͘Өڹ͢Δ
floatͷॏΈΛ1ݸஔ͘ϝϞϦ͕͋ͬͨΒ
uint8_tͷॏΈΛ4ݸஔ͍ͨ΄͏͕ྑ͍

View Slide

VkDeviceMemory
VkBuffer
0x8000000
Buffer device address
GPUͷϝϞϦ্ʹ͋ΔόοϑΝͷ
GPU಺Ͱͷઌ಄ΞυϨεΛऔಘ͢Δ
༻్1: σόοά৘ใʹΞυϨεΛࡌͤΔ

View Slide

#version 450
...
#extension GL_EXT_buffer_reference : enable
layout(buffer_reference) buffer node_t;
layout(buffer_reference, std430, buffer_reference_align = 16) buffer node_t
{
int value;
node_t next;
};
layout(std430) buffer uniforms_t {
node_t root;
} uniforms;
void main() {
node_t node = uniforms.root;
node = b.next.next;
...
} Buffer device address
༻్2: όοϑΝͷσʔλʹ
ଞͷόοϑΝͷΞυϨεΛॻ͘
GPU্ͰḷΕΔlinked listΛ࡞ΕΔ
GLSLͷbuffer_reference֦ுΛ࢖ͬͯಡΉ

View Slide

#version 450
...
layout(binding = 1) uniform sampler2D tex1;
...
int main() {
vec4 value = texture2D( tex5, tex_coord );
}
γΣʔμʹ౉͢
Ϧιʔε͕૿͑ͯ͘Δͱ
ਏ͍ίʔυ͕Ͱ͖Δ

View Slide

#version 450
...
layout(binding = 1) uniform sampler2D tex[];
...
int main() {
vec4 value = texture2D( tex[ 4 ], tex_coord );
}
σεΫϦϓλͷ഑ྻ
Λ࡞ΕΔΑ͏ʹ͢Δ
Descriptor Indexing

View Slide

#version 450
...
layout(binding = 1) uniform sampler2D tex[];
...
int main() {
}
σεΫϦϓλͷ഑ྻ
Λ࡞ΕΔΑ͏ʹ͢Δ
Descriptor Indexing
γΣʔμ͕৮Βͳ͍σεΫϦϓλ͸
࣮ࡍͷϦιʔεʹ݁ͼ͍͍ͭͯͳͯ͘΋ྑ͍
σεΫϦϓληοτͷཁ݅ͷ؇࿨
ίϚϯυόοϑΝͷه࿥தͰ΋
ࠓ৮ͬͯͳ͍σεΫϦϓλ͸ߋ৽ͯ͠Α͍

View Slide

int main() {
}
Λ࡞ΕΔΑ͏ʹ͢Δ
Descriptor Indexing
γΣʔμ͕৮Βͳ͍σεΫϦϓλ͸
࣮ࡍͷϦιʔεʹ݁ͼ͍͍ͭͯͳͯ͘΋ྑ͍
σεΫϦϓληοτͷཁ݅ͷ؇࿨
ίϚϯυόοϑΝͷه࿥தͰ΋
ࠓ৮ͬͯͳ͍σεΫϦϓλ͸ߋ৽ͯ͠Α͍
ͱΓ͋͑ͣڊେͳσεΫϦϓληοτΛ࡞͓͍ͬͯͯ
ඞཁʹԠͯ͡ඞཁͳཁૉʹϦιʔεΛηοτ͢Δӡ༻͕Մೳʹ

View Slide

ϑϨʔϜόοϑΝ
VkDeviceMemory
VkImage
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
VkDeviceMemory
VkImage
VkResult vkCreateFramebuffer(
VkDevice device,
const VkFramebufferCreateInfo* pCreateInfo,
VkFramebuffer* pFramebuffer
);
typedef struct VkFramebufferCreateInfo {
const void* pNext;
VkFramebufferCreateFlags flags;
VkRenderPass renderPass;
uint32_t width;
uint32_t height;
uint32_t layers;
} VkFramebufferCreateInfo;
࢖͏Πϝʔδͷ
Ϗϡʔͷ഑ྻ
ϑϨʔϜόοϑΝΑΓઌʹ
Πϝʔδ͕ཁΔ

View Slide

sType;
pNext;
Flags flags;
renderPass;
attachmentCount;
pAttachments;
width;
height;
layers;
Info;
NULL
typedef struct VkFramebufferAttachmentsCreateInfo {
const void* pNext;
uint32_t attachmentImageInfoCount;
const VkFramebufferAttachmentImageInfo* pAttachmentImageInfos;
} VkFramebufferAttachmentsCreateInfo;
VK_FRAMEBUFFER_CREATE_IMAGELESS_BIT_KHR
༁:͋ͱͰ
typedef struct VkFramebufferAttachmentImageInfo {
const void* pNext;
uint32_t width;
uint32_t height;
uint32_t layerCount;
uint32_t viewFormatCount;
const VkFormat* pViewFormats;
} VkFramebufferAttachmentImageInfo;
༁:͜ΜͳΠϝʔδϏϡʔ͕
෇͘༧ఆ
Imageless framebuffer

View Slide

NULL
༁:͋ͱͰ
typedef struct VkFramebufferAttachmentImageInfo {
const void* pNext;
uint32_t width;
uint32_t height;
uint32_t layerCount;
uint32_t viewFormatCount;
const VkFormat* pViewFormats;
} VkFramebufferAttachmentImageInfo;
༁:͜ΜͳΠϝʔδϏϡʔ͕
෇͘༧ఆ
Imageless framebuffer
typedef struct VkRenderPassAttachmentBeginInfo {
const void* pNext;
} VkRenderPassAttachmentBeginInfo;
࢖͏Πϝʔδͷ
Ϗϡʔͷ഑ྻ
ϨϯμʔύεΛΩϡʔʹ౤͛Δͱ͖ʹ͜ΕΛ෇͚ͯ
࢖͏ΠϝʔδϏϡʔΛܾఆ

View Slide

ϑϨʔϜόοϑΝ
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
৭͕ೖͬͯΔ ਂ౓ͱεςϯγϧ͕
ೖͬͯΔ
VulkanͰ͸ਂ౓ͱεςϯγϧ͸ಉ͡Πϝʔδʹه࿥͢Δ
Ұൠతͳਂ౓͕24bitɺεςϯγϧ͸8bitͰे෼ͳͷͰ
྆ऀΛ͚ͬͭͯ͘32bitʹ͢Δͱऩ·Γ͕ྑ͍

View Slide

VkDeviceMemory
VkImage
ਂ౓ͱεςϯγϧ͕
ೖͬͯΔ
͜Ε͸࣮ࡍʹ͸ґଘ͕ͳ͍σʔλ΁ͷґଘؔ܎Λੜͤ͡͞Δ
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
ό
Ϧ
Ξ
ਂ౓͔͍͠Βͳ͍Μ͚ͩͲ
͍ͬͭͯ͘Δ͔Β
྆ํʹґଘ͢Δ͔͠ͳ͍

View Slide

VkDeviceMemory
VkImage
͜Ε͸࣮ࡍʹ͸ґଘ͕ͳ͍σʔλ΁ͷґଘؔ܎Λੜͤ͡͞Δ
FS
Color Blend
typedef struct VkAttachmentDescriptionStencilLayout {
void* pNext;
VkImageLayout stencilInitialLayout;
VkImageLayout stencilFinalLayout;
} VkAttachmentDescriptionStencilLayout;
ਂ౓εςϯγϧͷΠϝʔδͷ͏ͪ
ͲͪΒ͔ҰํʹͷΈґଘ͕͋ΔࣄΛ໌ࣔͰ͖ΔΑ͏ʹ͢Δ
Separate Depth Stencil Layouts

View Slide

#version 450
#extension GL_ARB_gpu_shader_int64 : enable
#extension GL_EXT_shader_atomic_int64 : enable
...
void main() {
uint64_t result = atomicCompSwap( data, 0, 1 );
...
}
ʮdataʹஔ͔Εͨ஋͕0ͩͬͨΒ1ʹ͢ΔʯΛෆՄ෼ʹߦ͏
GPU͕αϙʔτ͍ͯ͠Δ৔߹
͜ͷΑ͏ͳ64bit੔਺ͷAtomicԋࢉΛγΣʔμͰ࢖͑ΔΑ͏ʹͳΔ
Atomic 64bit

View Slide

#version 450
...
f16vec4 input_bufffer[];
};
f16vec4 output_buffer[];
};
...
void main() {
vec4 value = input_buffer[ gl_GlobalInvocationID.x ];
output_buffer[ gl_GlobalInvocationID.x ] = value * 2.0;
}
൒ਫ਼౓
൒ਫ਼౓
୯ਫ਼౓
Vulkan 1.1ͷ16bitετϨʔδ͸
16bitͰϝϞϦʹஔ͍ͯ32bitͰܭࢉͩͬͨ

View Slide

#version 450
...
f16vec4 input_bufffer[];
};
f16vec4 output_buffer[];
};
...
void main() {
f16vec4 value = input_buffer[ gl_GlobalInvocationID.x ];
output_buffer[ gl_GlobalInvocationID.x ] = value * 2.0;
}
൒ਫ਼౓
൒ਫ਼౓
൒ਫ਼౓
Float16 Int8
Vulkan 1.2Ͱ͸σόΠε͕αϙʔτ͍ͯ͠Δ৔߹
൒ਫ਼౓ͷ··ܭࢉ͕Ͱ͖Δ

View Slide

#version 450
...
uint8_t input_bufffer[];
};
uint8_t output_buffer[];
};
...
void main() {
uint8_t value = input_buffer[ gl_GlobalInvocationID.x ];
output_buffer[ gl_GlobalInvocationID.x ] = value * 2;
}
8bit੔਺
8bit੔਺
8bit੔਺
Float16 Int8
Vulkan 1.2Ͱ͸σόΠε͕αϙʔτ͍ͯ͠Δ৔߹
8bit੔਺ͷ··ܭࢉ͕Ͱ͖Δ

View Slide

ίϚϯυόοϑΝ
ηϚϑΥ
ίϚϯυόοϑΝ
ηϚϑΥ
ίϚϯυόοϑΝ
ηϚϑΥ
ίϚϯυόοϑΝ
ηϚϑΥ
ίϚϯυόοϑΝ
ผͷΩϡʔͷίϚϯυͱ
ಉظΛऔΔʹ͸
ಉظճ਺෼ͷηϚϑΥ͕ཁΔ
͜Εͱ
͜Εͱ
͜Εͱ
͋ͱ͜Ε΋

View Slide

ίϚϯυόοϑΝ
ηϚϑΥ
ίϚϯυόοϑΝ
ίϚϯυόοϑΝ
ίϚϯυόοϑΝ
ίϚϯυόοϑΝ
1ͭͷηϚϑΥΛΧ΢ϯτ͍ͯ͘͠
ηϚϑΥΛ+1
ηϚϑΥ͕1ʹͳͬͨΒ։࢝
ηϚϑΥΛ+1
ηϚϑΥΛ+1
ηϚϑΥΛ+1
ηϚϑΥΛ+1
ಉظՕॴ͕ଟ͍৔߹ʹ؅ཧָ͕
Timeline Semaphore

View Slide

ίϚϯυόοϑΝ
ηϚϑΥ
ίϚϯυόοϑΝ
ίϚϯυόοϑΝ
ઌߦ͢Δ3ͭͷίϚϯυόοϑΝͷ͏ͪ
2͕ͭ׬ྃͨ͠Β̐ͭ໨Λ౤ೖͯ͠ྑ͍
ηϚϑΥΛ+1
ηϚϑΥΛ+1
Timeline Semaphore
ίϚϯυόοϑΝ
ηϚϑΥΛ+1

View Slide

ඪ४ʹೖ͍ͬͯͳ͍ϗοτͳ֦ு

View Slide

VK_KHR_video_queue
ίϚϯυόοϑΝ
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkBuffer
͜ͷόοϑΝʹೖͬͨ
ಈըͷετϦʔϜΛ
σίʔυͯ͠
͜ͷΠϝʔδͷྻʹు͍ͯ
ಈըରԠΩϡʔ GPU͕උ͑Δ
ϋʔυ΢ΣΞಈըΤϯίʔμɾσίʔμΛ࢖͏

View Slide

ैདྷͷ
ΠϯλϥΫςΟϒͳ
3DάϥϑΟΫε͸
ؒ઀র໌Λແࢹ͢Δ

View Slide

ʹ͓͚Δؒ઀র໌Λܭࢉ͢Δʹ͸
ͷҐஔ͔Β͋Δํ޲΁৳ͼΔઢ෼ ͕
ͷҐஔͰ ଞͷ໘ͱަࠩ͢ΔࣄΛ
ൃݟ͠ͳ͚Ε͹ͳΒͳ͍
p
p v
q
p
q
v

View Slide

v ⋮
௖఺഑ྻ
͸
ઢ෼v
ͱަࠩ͠·͔͢?
௖఺഑ྻͷࡾ֯ܗΛ1ͭͮͭᢞΊΔΑΓ
ޮ཰ͷྑ͍൑ఆํ๏͕ͳ͍
ϦΞϧλΠϜͰ൑ఆͯ͠
Ͱ͖·ͤΜ!

View Slide

v
௖఺഑ྻ
͸
ઢ෼v
ͱަࠩ͠·͔͢?
ϦΞϧλΠϜͰ൑ఆͯ͠
Ͱ͖·͢
ࣄલʹม׵
໦ߏ଄
ϦΞϧλΠϜͰ
มܗʹ௥ैͯ͠ Ͱ͖·ͤΜ!
௖఺഑ྻΛ໦ߏ଄ʹม׵
൑ఆ͸Ͱ͖Δɺ͕

View Slide

ڞ༗ϝϞϦ L1Ωϟογϡ
RT Core
࠷ۙͷNVIDIAͷ
GPUʹࡌͬͯΔ
RT Core ௖఺഑ྻ͔Β
BVH(໦ߏ଄)Λ
ര଎Ͱ࡞Γ
ര଎Ͱઢ෼ͱͷ
ަࠩ൑ఆΛ͢Δ
ઐ༻ϋʔυ΢ΣΞ

View Slide

VK_KHR_acceleration_structure
VkDeviceMemory
VkAccelerationStructureKHR
͜ͷϝϞϦΛަࠩ൑ఆͷҝʹ
GPU͕ੜ੒ͨ͠໦ߏ଄ͷஔ͖৔ॴͱͯ͠࢖͍·͢
۩ମతͳϑΥʔϚοτ͸Vulkanʹ೚ͤ·͢
VkDeviceMemory
VkImage
VkDeviceMemory
VkBuffer
͜ΕͷࣄΛVulkanͰ͸Acceleration StructureͱݺͿ

View Slide

void vkCmdBuildAccelerationStructuresKHR(
uint32_t infoCount,
const VkAccelerationStructureBuildGeometryInfoKHR* pInfos,
const VkAccelerationStructureBuildRangeInfoKHR* const* ppBuildRangeInfos
); typedef struct VkAccelerationStructureBuildGeometryInfoKHR {
const void* pNext;
VkAccelerationStructureTypeKHR type;
VkBuildAccelerationStructureFlagsKHR flags;
VkBuildAccelerationStructureModeKHR mode;
VkAccelerationStructureKHR srcAccelerationStructure;
VkAccelerationStructureKHR dstAccelerationStructure;
uint32_t geometryCount;
const VkAccelerationStructureGeometryKHR* pGeometries;
const VkAccelerationStructureGeometryKHR* const* ppGeometries;
VkDeviceOrHostAddressKHR scratchData;
} VkAccelerationStructureBuildGeometryInfoKHR;
͜Εʹ
޲͔ͬͯ

View Slide

onStructureGeometryKHR* pGeometries;
onStructureGeometryKHR* const* ppGeometries;
essKHR scratchData;
ctureBuildGeometryInfoKHR;
typedef struct VkAccelerationStructureGeometryKHR {
const void* pNext;
VkGeometryTypeKHR geometryType;
VkAccelerationStructureGeometryDataKHR geometry;
VkGeometryFlagsKHR flags;
} VkAccelerationStructureGeometryKHR;
typedef union VkAccelerationStructureGeometryDataKHR {
VkAccelerationStructureGeometryTrianglesDataKHR triangles;
VkAccelerationStructureGeometryAabbsDataKHR aabbs;
VkAccelerationStructureGeometryInstancesDataKHR instances;
} VkAccelerationStructureGeometryDataKHR;

View Slide

uctureGeometryKHR;
n VkAccelerationStructureGeometryDataKHR {
tionStructureGeometryTrianglesDataKHR triangles;
tionStructureGeometryAabbsDataKHR aabbs;
tionStructureGeometryInstancesDataKHR instances;
tionStructureGeometryDataKHR;
typedef struct VkAccelerationStructureGeometryTrianglesDataKHR {
const void* pNext;
VkFormat vertexFormat;
VkDeviceOrHostAddressConstKHR vertexData;
VkDeviceSize vertexStride;
uint32_t maxVertex;
VkIndexType indexType;
VkDeviceOrHostAddressConstKHR indexData;
VkDeviceOrHostAddressConstKHR transformData;
} VkAccelerationStructureGeometryTrianglesDataKHR;
͜ͷΞυϨεʹ
ஔ͍ͯ͋Δ
௖఺഑ྻ͔Β
໦ߏ଄Λੜ੒͢ΔίϚϯυΛΩϡʔʹੵΉ

View Slide

uctureGeometryKHR;
n VkAccelerationStructureGeometryDataKHR {
tionStructureGeometryTrianglesDataKHR triangles;
tionStructureGeometryAabbsDataKHR aabbs;
tionStructureGeometryInstancesDataKHR instances;
tionStructureGeometryDataKHR;
typedef struct VkAccelerationStructureGeometryAabbsDataKHR {
const void* pNext;
VkDeviceOrHostAddressConstKHR data;
VkDeviceSize stride;
} VkAccelerationStructureGeometryAabbsDataKHR;
͜ͷΞυϨεʹ
ஔ͍ͯ͋Δ
AABBͷ഑ྻ͔Β
໘ͱͷަࠩͰ͸ͳ͘
AABBͱͷަࠩ൑ఆΛ͢Δ໦ߏ଄Λ࡞Δ͜ͱ΋Ͱ͖Δ

View Slide

#version 450
#extension GL_EXT_ray_query : enable
...
void main() {
rayQueryEXT ray_query;
rayQueryInitializeEXT(
ray_query,
acceleration_structure,
gl_RayFlagsTerminateOnFirstHitEXT,
cull_mask,
pos,
near,
direction,
far
);
while( rayQueryProceedEXT( ray_query ) ) {
if(
rayQueryGetIntersectionTypeEXT( ray_query, false ) ==
gl_RayQueryCandidateIntersectionTriangleEXT
) {
rayQueryConfirmIntersectionEXT( ray_query );
}
}
if(
rayQueryGetIntersectionTypeEXT( ray_query, true) ==
gl_RayQueryCommittedIntersectionNoneEXT
) {
...
}
}
VK_KHR_ray_query
͜ͷAcceleration StructureͰ
posͷҐஔ͔Βdirectionͷ޲͖ʹ
near͔Βfar·Ͱͷڑ཭ͷઢ෼͕
Կ͔ͱަࠩ͢Δ͔ௐ΂ͯ
ަࠩ͢Δࡾ֯ܗΛΈ͚ͭͨΒ
ःṭ෺͕͋Δͱ͖ͷॲཧ

View Slide

෺ମͷද໘͕׬શͳڸ໘Ͱͳ͍ݶΓ
෺ମͷද໘ʹ౰ͨͬͨޫ͸༷ʑͳํ޲ʹࢄΒ͹͍ͬͯ͘
ϨΠτϨʔγϯάͰ͸
෺ମͷද໘ʹͿ͔ͭΔͨͼʹ
σʔλͷฒྻ౓্͕͕͍ͬͯ͘

View Slide

ϨΠτϨʔγϯάͰ͸
෺ମͷද໘ʹͿ͔ͭΔͨͼʹ
σʔλͷฒྻ౓্͕͕͍ͬͯ͘
͜ΕΛطଘͷ
ύΠϓϥΠϯͰߦ͏
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
CS
ίϯϐϡʔτύΠϓϥΠϯ άϥϑΟΫεύΠϓϥΠϯ

View Slide

Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
CS
ίϯϐϡʔτύΠϓϥΠϯ άϥϑΟΫεύΠϓϥΠϯ
͜ΕΛطଘͷ
ύΠϓϥΠϯͰߦ͏
ͷ͸ແཧͦ͏ͩͬͨͷͰ৽͍͠ύΠϓϥΠϯ͕ੜ͑ͨ
RayGen Shader
Closest Hit Shader Miss Shader
ϨΠτϨʔγϯάύΠϓϥΠϯ
VK_KHR_ray_tracing_pipeline
Ray Query

View Slide

X Window System
Wayland Compositor
Windows DWM
etc.
Vulkan
ΞϓϦέʔγϣϯ
ίϯϙδλΛܦ༝͢ΔΦʔόʔϔου͕զຫͰ͖ͳ͍

View Slide

Windows DWM
Vulkan
ΞϓϦέʔγϣϯ
શը໘දࣔதͳΒΞϓϦέʔγϣϯଆʹ
σΟεϓϨΠ΁ͷग़ྗ಺༰Λ௚઀৮Βͤͯ΋ྑ͍ͷͰ͸
vkAcquireFullScreenExclusiveModeEXT
(༁:ը໘Λؙ͝ͱΑͤ͜)
VK_EXT_full_screen_exclusive

View Slide

͜͜ʹॻ͘ͱग़Δ
XΛىಈ͍ͯ͠ͳ͍Linux
Vulkan
ΞϓϦέʔγϣϯ
ͦ΋ͦ΋ίϯϙδλ͕ډͳ͍ͳΒ
ΞϓϦέʔγϣϯ͕σΟεϓϨΠͷ੍ޚΛѲͬͯྑ͍ͷͰ͸
ίϯϙδλ
ͲΜͳϞʔυͰදࣔͰ͖ΔσΟεϓϨΠ͕
͍ͭ͘ܨ͕͍ͬͯ·͔͢?
VK_KHR_display
σΟεϓϨΠ1

View Slide

͜͜ʹॻ͘ͱग़Δ
Vulkan
ΞϓϦέʔγϣϯ
LinuxͷKernel Mode Settingʹର͢Δബ͍ϥούʔ͕
Vulkanʹ௥Ճ͞ΕΔ
σΟεϓϨΠ1΁ͷग़ྗΛ[email protected] 24bitʹͯ͠
ͦ͜ʹॻͨ͘ΊͷεϫοϓνΣʔϯΛ࡞੒
VK_KHR_display_swapchain
εϫοϓνΣʔϯ
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
VkDeviceMemory
VkImage
σΟεϓϨΠ1

View Slide

ϝογϡͷڥք෦෼Ҏ֎Ͱ͸
ۙ๣ͷϐΫηϧͱࣅͨ৭ʹͳΔϐΫηϧ͕ଟ͍

View Slide

ࣄલʹڥք͕Ͳ͜ʹདྷΔ͔Θ͔Δ৔߹
ͦΕʹج͍ͮͯϑϥάϝϯτγΣʔμͷ࣮ߦΛؒҾ͖͍ͨ
Fragment Density Map

View Slide

≃
ؒҾ͍ͨ৔߹ શͯܭࢉͨ͠৔߹
VK_EXT_fragment_density_map

View Slide

VK_EXT_fragment_density_map
ਓؒͷࢹ֮͸ࢹ໺ͷத৺෦෼Ҏ֎͸ࡉ͔͍ྠֲΛଊ͍͑ͯͳ͍
ࢹઢΛ௥੻Ͱ͖ΔVRϔουηοτͰத৺෇͚ۙͩࡉ͔͘ඳ͖͍ͨ

View Slide

VK_KHR_fragment_shading_rate
MSAA΍SupersamplingͰ͸
ΞϯνΤΠϦΞεͷҝʹ1ϐΫηϧʹରͯ͠
ϑϥάϝϯτγΣʔμͷ࣮ߦ݁ՌΛෳ਺࣋ͭ
ڥք෦෼Ͱ͸༗ޮ͕ͩ
ͦΕҎ֎Ͱ͸ແବͳͷͰ
৔ॴʹΑͬͯݸ਺Λม͍͑ͨ

View Slide

Input Assembly
Vertex Shader
Tessellation
Geometry Shader
Rasterization
Fragment Shader
Color Blend
VK_EXT_transform_feedback
VkDeviceMemory
VkBufer
άϥϑΟΫεύΠϓϥΠϯΛ
δΦϝτϦγΣʔμ·ͰͰࢭΊͯ
δΦϝτϦγΣʔμͷग़ྗΛ
όοϑΝʹు͘
OpenGLʹ͸ඪ४ͰඋΘͬͯͨ΍ͭ

View Slide

Ϩϯμʔύε
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
"
Ϩϯμʔύε
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
#
Ϩϯμʔύε
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
$
Ϩϯμʔύε
Input Assembly
VS
TCS
Tessellation
TES
GS
Rasterization
FS
Color Blend
%
ϞόΠϧGPUͰͳ͍GPUͰ͸
ϨϯμʔύεΛ׆༻͢Δҙຯ͸͋·Γͳ͍ͷͰ
ύΠϓϥΠϯ͕1͚ͭͩͷϨϯμʔύε͕େྔʹͰ͖͕ͪ
ϨϯμʔύεΛ࡞Δͷ͕ΊΜͲ͍͘͞

View Slide

VK_KHR_dynamic_rendering
ϨϯμʔύεΛ
NULLͰ΋ྑ͘͢Δ
άϥϑΟΫεύΠϓϥΠϯ
࡞੒࣌

View Slide

VK_KHR_dynamic_rendering
void vkCmdBeginRenderingKHR(
VkRenderingInfoKHR* pRenderingInfo
);
void vkCmdEndRenderingKHR(
VkCommandBuffer commandBuffer
);
͔͜͜Βଈ੮Ͱ࡞ͬͨ
ϨϯμʔύεΛ࢖͏
͜͜·Ͱଈ੮Ͱ࡞ͬͨ
ϨϯμʔύεΛ࢖͏
த਎͕ύΠϓϥΠϯ1͚ͭͩͷϨϯμʔύεͳΒ
ϨϯμʔύεΛίϚϯυόοϑΝʹੵΉ࣌ʹ
ͦͷ৔Ͱ࡞ΕΔΑ͏ʹ͢Δ

View Slide

ٕज़ॻయ12Ͱ
࠷ۙͷVulkanͷ࿩Λ੝ΓࠐΜͩ
3DάϥϑΟΫεAPI
VulkanΛ
ग़དྷΔ͚ͩ
΍͘͞͠
ղઆ͢Δຊ
Version 3.0
ΛϦϦʔε༧ఆ
※ࠨͷը૾͸Version 2.0ͷ΋ͷͰ͢
ిࢠ൛ͷ1.0·ͨ͸2.0Λ͍࣋ͬͯΔ৔߹
ແྉͰΞοϓσʔτΛड͚ΒΕ·͢
※

View Slide

·ͱΊ
GPU͸୔ࢁͷϓϩηοα͕ࡌͬͨܭࢉػͩ
VulkanΛ࢖͑͹GPUͷҰ௨Γͷૢ࡞͕Ͱ͖Δ
Vulkan͸վྑ͕ଓ͚ΒΕ͍ͯΔ

View Slide

いまどきのVulkan

いまどきのVulkan

Fadis

More Decks by Fadis

Other Decks in Programming

Featured

Transcript