Enum CudaDeviceAttribute

Namespace

DotCompute.Backends.CUDA.Types.Native

Assembly

DotCompute.Backends.CUDA.dll

CUDA device attributes that can be queried via cudaDeviceGetAttribute

public enum CudaDeviceAttribute

Fields

AsyncEngineCount = 15

Device can possibly copy memory and execute a kernel concurrently

CanFlushRemoteWrites = 56

Device supports flushing denormalized values to zero

CanMapHostMemory = 19

Device can map host memory into CUDA address space

CanUse64BitStreamMemOps = 78

Device supports GPUDirect RDMA

CanUseHostPointerForRegisteredMem = 52

Device can use host memory as registered memory

CanUseStreamMemOps = 77

Device supports caching of memory accesses to remote GPUs

CanUseStreamWaitValueNor = 79

Device supports IPC events

ClockRate = 13

Typical clock frequency in kilohertz

ComputeCapabilityMajor = 75

Major compute capability version number

ComputeCapabilityMinor = 76

Minor compute capability version number

ComputeMode = 20

Compute mode (see CudaComputeMode)

ComputePreemptionSupported = 51

Device supports Compute Preemption

ConcurrentKernels = 31

Device can execute multiple kernels concurrently

ConcurrentManagedAccess = 50

Device can coherently access managed memory concurrently with the CPU

CooperativeLaunch = 53

Device supports cooperative kernels

CooperativeMultiDeviceLaunch = 54

Device supports cooperative kernels across multiple devices

CooperativeMultiDeviceUnmatchedBlockDim = 82

Device supports shared memory configurations

CooperativeMultiDeviceUnmatchedFunc = 80

Device supports launching cooperative kernels via cudaLaunchCooperativeKernelMultiDevice

CooperativeMultiDeviceUnmatchedGridDim = 81

Device supports shared memory carveout configurations

CooperativeMultiDeviceUnmatchedSharedMem = 83

Device supports async memory operations

DirectManagedMemAccessFromHost = 59

Host can directly access managed memory on the device without migration

ECCEnabled = 32

Device has ECC support enabled

GlobalL1CacheSupported = 41

Device supports caching globals in L1

GlobalMemoryBusWidth = 37

Global memory bus width in bits

HostRegisterReadOnlySupported = 57

Device supports host memory registration with read-only flag

HostRegisterSupported = 48

Device supports host memory registration

Integrated = 17

Device is integrated with host memory

L2CacheSize = 38

Size of L2 cache in bytes

LocalL1CacheSupported = 42

Device supports caching locals in L1

ManagedMemory = 45

Device can allocate managed memory

MaxBlockDimX = 2

Maximum x-dimension of a block

MaxBlockDimY = 3

Maximum y-dimension of a block

MaxBlockDimZ = 4

Maximum z-dimension of a block

MaxBlocksPerMultiprocessor = 49

Maximum number of blocks per multiprocessor

MaxGridDimX = 5

Maximum x-dimension of a grid

MaxGridDimY = 6

Maximum y-dimension of a grid

MaxGridDimZ = 7

Maximum z-dimension of a grid

MaxPitch = 11

Maximum pitch in bytes allowed by memory copies

MaxRegistersPerBlock = 12

Maximum number of 32-bit registers available per block

MaxRegistersPerMultiprocessor = 44

Maximum number of 32-bit registers available per multiprocessor

MaxSharedMemoryPerBlock = 8

Maximum shared memory available per block in bytes

MaxSharedMemoryPerBlockOptin = 55

Maximum shared memory available per block for dynamic allocation in bytes

MaxSharedMemoryPerMultiprocessor = 43

Maximum shared memory available per multiprocessor in bytes

MaxThreadsPerBlock = 1

Maximum number of threads per block

MaxThreadsPerMultiProcessor = 39

Maximum resident threads per multiprocessor

MemoryClockRate = 36

Peak memory clock frequency in kilohertz

MemoryPoolsSupported = 115

Device supports stream ordered memory allocator

MultiGpuBoard = 46

Device is on a multi-GPU board

MultiGpuBoardGroupID = 47

Unique ID for a group of devices on the same multi-GPU board

PCIBusID = 33

PCI bus ID of the device

PCIDeviceID = 34

PCI device ID of the device

PageableMemoryAccessUsesHostPageTables = 58

Shared memory regions can be allocated atomically

StreamPrioritiesSupported = 40

Device supports stream priorities

TccDriver = 35

Device is using TCC driver model

TextureAlignment = 14

Alignment requirement for textures

TotalConstantMemory = 9

Memory available on device for constant variables in a CUDA kernel in bytes

UnifiedAddressing = 18

Device supports unified virtual addressing with the host

WarpSize = 10

Warp size in threads