hipDeviceProp_t Struct Reference#
#include <hip_runtime_api.h>
Data Fields | |
| char | name [256] |
| Device name. | |
| hipUUID | uuid |
| UUID of a device. | |
| char | luid [8] |
| 8-byte unique identifier. Only valid on windows | |
| unsigned int | luidDeviceNodeMask |
| LUID node mask. | |
| size_t | totalGlobalMem |
| Size of global memory region (in bytes). | |
| size_t | sharedMemPerBlock |
| Size of shared memory per block (in bytes). | |
| int | regsPerBlock |
| Registers per block. | |
| int | warpSize |
| Warp size. | |
| size_t | memPitch |
| int | maxThreadsPerBlock |
| Max work items per work group or workgroup max size. | |
| int | maxThreadsDim [3] |
| Max number of threads in each dimension (XYZ) of a block. | |
| int | maxGridSize [3] |
| Max grid dimensions (XYZ). | |
| int | clockRate |
| Max clock frequency of the multiProcessors in khz. | |
| size_t | totalConstMem |
| int | major |
| int | minor |
| size_t | textureAlignment |
| Alignment requirement for textures. | |
| size_t | texturePitchAlignment |
| Pitch alignment requirement for texture references bound to. | |
| int | deviceOverlap |
| Deprecated. Use asyncEngineCount instead. | |
| int | multiProcessorCount |
| int | kernelExecTimeoutEnabled |
| Run time limit for kernels executed on the device. | |
| int | integrated |
| APU vs dGPU. | |
| int | canMapHostMemory |
| Check whether HIP can map host memory. | |
| int | computeMode |
| Compute mode. | |
| int | maxTexture1D |
| Maximum number of elements in 1D images. | |
| int | maxTexture1DMipmap |
| Maximum 1D mipmap texture size. | |
| int | maxTexture1DLinear |
| Maximum size for 1D textures bound to linear memory. | |
| int | maxTexture2D [2] |
| Maximum dimensions (width, height) of 2D images, in image elements. | |
| int | maxTexture2DMipmap [2] |
| Maximum number of elements in 2D array mipmap of images. | |
| int | maxTexture2DLinear [3] |
| Maximum 2D tex dimensions if tex are bound to pitched memory. | |
| int | maxTexture2DGather [2] |
| Maximum 2D tex dimensions if gather has to be performed. | |
| int | maxTexture3D [3] |
| int | maxTexture3DAlt [3] |
| Maximum alternate 3D texture dims. | |
| int | maxTextureCubemap |
| Maximum cubemap texture dims. | |
| int | maxTexture1DLayered [2] |
| Maximum number of elements in 1D array images. | |
| int | maxTexture2DLayered [3] |
| Maximum number of elements in 2D array images. | |
| int | maxTextureCubemapLayered [2] |
| Maximum cubemaps layered texture dims. | |
| int | maxSurface1D |
| Maximum 1D surface size. | |
| int | maxSurface2D [2] |
| Maximum 2D surface size. | |
| int | maxSurface3D [3] |
| Maximum 3D surface size. | |
| int | maxSurface1DLayered [2] |
| Maximum 1D layered surface size. | |
| int | maxSurface2DLayered [3] |
| Maximum 2D layared surface size. | |
| int | maxSurfaceCubemap |
| Maximum cubemap surface size. | |
| int | maxSurfaceCubemapLayered [2] |
| Maximum cubemap layered surface size. | |
| size_t | surfaceAlignment |
| Alignment requirement for surface. | |
| int | concurrentKernels |
| Device can possibly execute multiple kernels concurrently. | |
| int | ECCEnabled |
| Device has ECC support enabled. | |
| int | pciBusID |
| PCI Bus ID. | |
| int | pciDeviceID |
| PCI Device ID. | |
| int | pciDomainID |
| PCI Domain ID. | |
| int | tccDriver |
| 1:If device is Tesla device using TCC driver, else 0 | |
| int | asyncEngineCount |
| Number of async engines. | |
| int | unifiedAddressing |
| Does device and host share unified address space. | |
| int | memoryClockRate |
| Max global memory clock frequency in khz. | |
| int | memoryBusWidth |
| Global memory bus width in bits. | |
| int | l2CacheSize |
| L2 cache size. | |
| int | persistingL2CacheMaxSize |
| Device's max L2 persisting lines in bytes. | |
| int | maxThreadsPerMultiProcessor |
| Maximum resident threads per multi-processor. | |
| int | streamPrioritiesSupported |
| Device supports stream priority. | |
| int | globalL1CacheSupported |
| Indicates globals are cached in L1. | |
| int | localL1CacheSupported |
| Locals are cahced in L1. | |
| size_t | sharedMemPerMultiprocessor |
| Amount of shared memory available per multiprocessor. | |
| int | regsPerMultiprocessor |
| registers available per multiprocessor | |
| int | managedMemory |
| Device supports allocating managed memory on this system. | |
| int | isMultiGpuBoard |
| 1 if device is on a multi-GPU board, 0 if not. | |
| int | multiGpuBoardGroupID |
| Unique identifier for a group of devices on same multiboard GPU. | |
| int | hostNativeAtomicSupported |
| Link between host and device supports native atomics. | |
| int | singleToDoublePrecisionPerfRatio |
| Deprecated. CUDA only. | |
| int | pageableMemoryAccess |
| int | concurrentManagedAccess |
| int | computePreemptionSupported |
| Is compute preemption supported on the device. | |
| int | canUseHostPointerForRegisteredMem |
| int | cooperativeLaunch |
| HIP device supports cooperative launch. | |
| int | cooperativeMultiDeviceLaunch |
| size_t | sharedMemPerBlockOptin |
| Per device m ax shared mem per block usable by special opt in. | |
| int | pageableMemoryAccessUsesHostPageTables |
| int | directManagedMemAccessFromHost |
| int | maxBlocksPerMultiProcessor |
| Max number of blocks on CU. | |
| int | accessPolicyMaxWindowSize |
| Max value of access policy window. | |
| size_t | reservedSharedMemPerBlock |
| Shared memory reserved by driver per block. | |
| int | hostRegisterSupported |
| Device supports hipHostRegister. | |
| int | sparseHipArraySupported |
| Indicates if device supports sparse hip arrays. | |
| int | hostRegisterReadOnlySupported |
| int | timelineSemaphoreInteropSupported |
| Indicates external timeline semaphore support. | |
| int | memoryPoolsSupported |
| Indicates if device supports hipMallocAsync and hipMemPool APIs. | |
| int | gpuDirectRDMASupported |
| Indicates device support of RDMA APIs. | |
| unsigned int | gpuDirectRDMAFlushWritesOptions |
| int | gpuDirectRDMAWritesOrdering |
| value of hipGPUDirectRDMAWritesOrdering | |
| unsigned int | memoryPoolSupportedHandleTypes |
| Bitmask of handle types support with mempool based IPC. | |
| int | deferredMappingHipArraySupported |
| int | ipcEventSupported |
| Device supports IPC events. | |
| int | clusterLaunch |
| Device supports cluster launch. | |
| int | unifiedFunctionPointers |
| Indicates device supports unified function pointers. | |
| int | reserved [63] |
| CUDA Reserved. | |
| int | hipReserved [32] |
| Reserved for adding new entries for HIP/CUDA. | |
| char | gcnArchName [256] |
| AMD GCN Arch Name. HIP Only. | |
| size_t | maxSharedMemoryPerMultiProcessor |
| Maximum Shared Memory Per CU. HIP Only. | |
| int | clockInstructionRate |
| hipDeviceArch_t | arch |
| Architectural feature flags. New for HIP. | |
| unsigned int * | hdpMemFlushCntl |
| Addres of HDP_MEM_COHERENCY_FLUSH_CNTL register. | |
| unsigned int * | hdpRegFlushCntl |
| Addres of HDP_REG_COHERENCY_FLUSH_CNTL register. | |
| int | cooperativeMultiDeviceUnmatchedFunc |
| int | cooperativeMultiDeviceUnmatchedGridDim |
| int | cooperativeMultiDeviceUnmatchedBlockDim |
| int | cooperativeMultiDeviceUnmatchedSharedMem |
| int | isLargeBar |
| 1: if it is a large PCI bar device, else 0 | |
| int | asicRevision |
| Revision of the GPU in this device. | |
Detailed Description
hipDeviceProp
Field Documentation
◆ accessPolicyMaxWindowSize
| int hipDeviceProp_t::accessPolicyMaxWindowSize |
Max value of access policy window.
◆ arch
| hipDeviceArch_t hipDeviceProp_t::arch |
Architectural feature flags. New for HIP.
◆ asicRevision
| int hipDeviceProp_t::asicRevision |
Revision of the GPU in this device.
◆ asyncEngineCount
| int hipDeviceProp_t::asyncEngineCount |
Number of async engines.
◆ canMapHostMemory
| int hipDeviceProp_t::canMapHostMemory |
Check whether HIP can map host memory.
◆ canUseHostPointerForRegisteredMem
| int hipDeviceProp_t::canUseHostPointerForRegisteredMem |
Device can access host registered memory with same address as the host
◆ clockInstructionRate
| int hipDeviceProp_t::clockInstructionRate |
Frequency in khz of the timer used by the device-side "clock*" instructions. New for HIP.
◆ clockRate
| int hipDeviceProp_t::clockRate |
Max clock frequency of the multiProcessors in khz.
◆ clusterLaunch
| int hipDeviceProp_t::clusterLaunch |
Device supports cluster launch.
◆ computeMode
| int hipDeviceProp_t::computeMode |
Compute mode.
◆ computePreemptionSupported
| int hipDeviceProp_t::computePreemptionSupported |
Is compute preemption supported on the device.
◆ concurrentKernels
| int hipDeviceProp_t::concurrentKernels |
Device can possibly execute multiple kernels concurrently.
◆ concurrentManagedAccess
| int hipDeviceProp_t::concurrentManagedAccess |
Device can coherently access managed memory concurrently with the CPU
◆ cooperativeLaunch
| int hipDeviceProp_t::cooperativeLaunch |
HIP device supports cooperative launch.
◆ cooperativeMultiDeviceLaunch
| int hipDeviceProp_t::cooperativeMultiDeviceLaunch |
HIP device supports cooperative launch on multiple devices
◆ cooperativeMultiDeviceUnmatchedBlockDim
| int hipDeviceProp_t::cooperativeMultiDeviceUnmatchedBlockDim |
HIP device supports cooperative launch on multiple devices with unmatched block dimensions
◆ cooperativeMultiDeviceUnmatchedFunc
| int hipDeviceProp_t::cooperativeMultiDeviceUnmatchedFunc |
HIP device supports cooperative launch on multiple devices with unmatched functions
◆ cooperativeMultiDeviceUnmatchedGridDim
| int hipDeviceProp_t::cooperativeMultiDeviceUnmatchedGridDim |
HIP device supports cooperative launch on multiple devices with unmatched grid dimensions
◆ cooperativeMultiDeviceUnmatchedSharedMem
| int hipDeviceProp_t::cooperativeMultiDeviceUnmatchedSharedMem |
HIP device supports cooperative launch on multiple devices with unmatched shared memories
◆ deferredMappingHipArraySupported
| int hipDeviceProp_t::deferredMappingHipArraySupported |
Device supports deferred mapping HIP arrays and HIP mipmapped arrays
◆ deviceOverlap
| int hipDeviceProp_t::deviceOverlap |
Deprecated. Use asyncEngineCount instead.
◆ directManagedMemAccessFromHost
| int hipDeviceProp_t::directManagedMemAccessFromHost |
Host can directly access managed memory on the device without migration
◆ ECCEnabled
| int hipDeviceProp_t::ECCEnabled |
Device has ECC support enabled.
◆ gcnArchName
| char hipDeviceProp_t::gcnArchName[256] |
AMD GCN Arch Name. HIP Only.
◆ globalL1CacheSupported
| int hipDeviceProp_t::globalL1CacheSupported |
Indicates globals are cached in L1.
◆ gpuDirectRDMAFlushWritesOptions
| unsigned int hipDeviceProp_t::gpuDirectRDMAFlushWritesOptions |
Bitmask to be interpreted according to hipFlushGPUDirectRDMAWritesOptions
◆ gpuDirectRDMASupported
| int hipDeviceProp_t::gpuDirectRDMASupported |
Indicates device support of RDMA APIs.
◆ gpuDirectRDMAWritesOrdering
| int hipDeviceProp_t::gpuDirectRDMAWritesOrdering |
value of hipGPUDirectRDMAWritesOrdering
◆ hdpMemFlushCntl
| unsigned int* hipDeviceProp_t::hdpMemFlushCntl |
Addres of HDP_MEM_COHERENCY_FLUSH_CNTL register.
◆ hdpRegFlushCntl
| unsigned int* hipDeviceProp_t::hdpRegFlushCntl |
Addres of HDP_REG_COHERENCY_FLUSH_CNTL register.
◆ hipReserved
| int hipDeviceProp_t::hipReserved[32] |
Reserved for adding new entries for HIP/CUDA.
◆ hostNativeAtomicSupported
| int hipDeviceProp_t::hostNativeAtomicSupported |
Link between host and device supports native atomics.
◆ hostRegisterReadOnlySupported
| int hipDeviceProp_t::hostRegisterReadOnlySupported |
Device supports using the hipHostRegisterReadOnly flag with hipHostRegistger
◆ hostRegisterSupported
| int hipDeviceProp_t::hostRegisterSupported |
Device supports hipHostRegister.
◆ integrated
| int hipDeviceProp_t::integrated |
APU vs dGPU.
◆ ipcEventSupported
| int hipDeviceProp_t::ipcEventSupported |
Device supports IPC events.
◆ isLargeBar
| int hipDeviceProp_t::isLargeBar |
1: if it is a large PCI bar device, else 0
◆ isMultiGpuBoard
| int hipDeviceProp_t::isMultiGpuBoard |
1 if device is on a multi-GPU board, 0 if not.
◆ kernelExecTimeoutEnabled
| int hipDeviceProp_t::kernelExecTimeoutEnabled |
Run time limit for kernels executed on the device.
◆ l2CacheSize
| int hipDeviceProp_t::l2CacheSize |
L2 cache size.
◆ localL1CacheSupported
| int hipDeviceProp_t::localL1CacheSupported |
Locals are cahced in L1.
◆ luid
| char hipDeviceProp_t::luid[8] |
8-byte unique identifier. Only valid on windows
◆ luidDeviceNodeMask
| unsigned int hipDeviceProp_t::luidDeviceNodeMask |
LUID node mask.
◆ major
| int hipDeviceProp_t::major |
Major compute capability version. This indicates the core instruction set of the GPU architecture. For example, a value of 11 would correspond to Navi III (RDNA3). See the arch feature flags for portable ways to query feature caps.
◆ managedMemory
| int hipDeviceProp_t::managedMemory |
Device supports allocating managed memory on this system.
◆ maxBlocksPerMultiProcessor
| int hipDeviceProp_t::maxBlocksPerMultiProcessor |
Max number of blocks on CU.
◆ maxGridSize
| int hipDeviceProp_t::maxGridSize[3] |
Max grid dimensions (XYZ).
◆ maxSharedMemoryPerMultiProcessor
| size_t hipDeviceProp_t::maxSharedMemoryPerMultiProcessor |
Maximum Shared Memory Per CU. HIP Only.
◆ maxSurface1D
| int hipDeviceProp_t::maxSurface1D |
Maximum 1D surface size.
◆ maxSurface1DLayered
| int hipDeviceProp_t::maxSurface1DLayered[2] |
Maximum 1D layered surface size.
◆ maxSurface2D
| int hipDeviceProp_t::maxSurface2D[2] |
Maximum 2D surface size.
◆ maxSurface2DLayered
| int hipDeviceProp_t::maxSurface2DLayered[3] |
Maximum 2D layared surface size.
◆ maxSurface3D
| int hipDeviceProp_t::maxSurface3D[3] |
Maximum 3D surface size.
◆ maxSurfaceCubemap
| int hipDeviceProp_t::maxSurfaceCubemap |
Maximum cubemap surface size.
◆ maxSurfaceCubemapLayered
| int hipDeviceProp_t::maxSurfaceCubemapLayered[2] |
Maximum cubemap layered surface size.
◆ maxTexture1D
| int hipDeviceProp_t::maxTexture1D |
Maximum number of elements in 1D images.
◆ maxTexture1DLayered
| int hipDeviceProp_t::maxTexture1DLayered[2] |
Maximum number of elements in 1D array images.
◆ maxTexture1DLinear
| int hipDeviceProp_t::maxTexture1DLinear |
Maximum size for 1D textures bound to linear memory.
◆ maxTexture1DMipmap
| int hipDeviceProp_t::maxTexture1DMipmap |
Maximum 1D mipmap texture size.
◆ maxTexture2D
| int hipDeviceProp_t::maxTexture2D[2] |
Maximum dimensions (width, height) of 2D images, in image elements.
◆ maxTexture2DGather
| int hipDeviceProp_t::maxTexture2DGather[2] |
Maximum 2D tex dimensions if gather has to be performed.
◆ maxTexture2DLayered
| int hipDeviceProp_t::maxTexture2DLayered[3] |
Maximum number of elements in 2D array images.
◆ maxTexture2DLinear
| int hipDeviceProp_t::maxTexture2DLinear[3] |
Maximum 2D tex dimensions if tex are bound to pitched memory.
◆ maxTexture2DMipmap
| int hipDeviceProp_t::maxTexture2DMipmap[2] |
Maximum number of elements in 2D array mipmap of images.
◆ maxTexture3D
| int hipDeviceProp_t::maxTexture3D[3] |
Maximum dimensions (width, height, depth) of 3D images, in image elements
◆ maxTexture3DAlt
| int hipDeviceProp_t::maxTexture3DAlt[3] |
Maximum alternate 3D texture dims.
◆ maxTextureCubemap
| int hipDeviceProp_t::maxTextureCubemap |
Maximum cubemap texture dims.
◆ maxTextureCubemapLayered
| int hipDeviceProp_t::maxTextureCubemapLayered[2] |
Maximum cubemaps layered texture dims.
◆ maxThreadsDim
| int hipDeviceProp_t::maxThreadsDim[3] |
Max number of threads in each dimension (XYZ) of a block.
◆ maxThreadsPerBlock
| int hipDeviceProp_t::maxThreadsPerBlock |
Max work items per work group or workgroup max size.
◆ maxThreadsPerMultiProcessor
| int hipDeviceProp_t::maxThreadsPerMultiProcessor |
Maximum resident threads per multi-processor.
◆ memoryBusWidth
| int hipDeviceProp_t::memoryBusWidth |
Global memory bus width in bits.
◆ memoryClockRate
| int hipDeviceProp_t::memoryClockRate |
Max global memory clock frequency in khz.
◆ memoryPoolsSupported
| int hipDeviceProp_t::memoryPoolsSupported |
Indicates if device supports hipMallocAsync and hipMemPool APIs.
◆ memoryPoolSupportedHandleTypes
| unsigned int hipDeviceProp_t::memoryPoolSupportedHandleTypes |
Bitmask of handle types support with mempool based IPC.
◆ memPitch
| size_t hipDeviceProp_t::memPitch |
Maximum pitch in bytes allowed by memory copies pitched memory
◆ minor
| int hipDeviceProp_t::minor |
Minor compute capability version. This indicates a particular configuration, feature set, or variation within the group represented by the major compute capability version. For example, different models within the same major version might have varying levels of support for certain features or optimizations. See the arch feature flags for portable ways to query feature caps.
◆ multiGpuBoardGroupID
| int hipDeviceProp_t::multiGpuBoardGroupID |
Unique identifier for a group of devices on same multiboard GPU.
◆ multiProcessorCount
| int hipDeviceProp_t::multiProcessorCount |
Number of multi-processors. When the GPU works in Compute Unit (CU) mode, this value equals the number of CUs; when in Workgroup Processor (WGP) mode, this value equels half of CUs, because a single WGP contains two CUs.
◆ name
| char hipDeviceProp_t::name[256] |
Device name.
◆ pageableMemoryAccess
| int hipDeviceProp_t::pageableMemoryAccess |
Device supports coherently accessing pageable memory without calling hipHostRegister on it
◆ pageableMemoryAccessUsesHostPageTables
| int hipDeviceProp_t::pageableMemoryAccessUsesHostPageTables |
Device accesses pageable memory via the host's page tables
◆ pciBusID
| int hipDeviceProp_t::pciBusID |
PCI Bus ID.
◆ pciDeviceID
| int hipDeviceProp_t::pciDeviceID |
PCI Device ID.
◆ pciDomainID
| int hipDeviceProp_t::pciDomainID |
PCI Domain ID.
◆ persistingL2CacheMaxSize
| int hipDeviceProp_t::persistingL2CacheMaxSize |
Device's max L2 persisting lines in bytes.
◆ regsPerBlock
| int hipDeviceProp_t::regsPerBlock |
Registers per block.
◆ regsPerMultiprocessor
| int hipDeviceProp_t::regsPerMultiprocessor |
registers available per multiprocessor
◆ reserved
| int hipDeviceProp_t::reserved[63] |
CUDA Reserved.
◆ reservedSharedMemPerBlock
| size_t hipDeviceProp_t::reservedSharedMemPerBlock |
Shared memory reserved by driver per block.
◆ sharedMemPerBlock
| size_t hipDeviceProp_t::sharedMemPerBlock |
Size of shared memory per block (in bytes).
◆ sharedMemPerBlockOptin
| size_t hipDeviceProp_t::sharedMemPerBlockOptin |
Per device m ax shared mem per block usable by special opt in.
◆ sharedMemPerMultiprocessor
| size_t hipDeviceProp_t::sharedMemPerMultiprocessor |
Amount of shared memory available per multiprocessor.
◆ singleToDoublePrecisionPerfRatio
| int hipDeviceProp_t::singleToDoublePrecisionPerfRatio |
Deprecated. CUDA only.
◆ sparseHipArraySupported
| int hipDeviceProp_t::sparseHipArraySupported |
Indicates if device supports sparse hip arrays.
◆ streamPrioritiesSupported
| int hipDeviceProp_t::streamPrioritiesSupported |
Device supports stream priority.
◆ surfaceAlignment
| size_t hipDeviceProp_t::surfaceAlignment |
Alignment requirement for surface.
◆ tccDriver
| int hipDeviceProp_t::tccDriver |
1:If device is Tesla device using TCC driver, else 0
◆ textureAlignment
| size_t hipDeviceProp_t::textureAlignment |
Alignment requirement for textures.
◆ texturePitchAlignment
| size_t hipDeviceProp_t::texturePitchAlignment |
Pitch alignment requirement for texture references bound to.
◆ timelineSemaphoreInteropSupported
| int hipDeviceProp_t::timelineSemaphoreInteropSupported |
Indicates external timeline semaphore support.
◆ totalConstMem
| size_t hipDeviceProp_t::totalConstMem |
Size of shared constant memory region on the device (in bytes).
◆ totalGlobalMem
| size_t hipDeviceProp_t::totalGlobalMem |
Size of global memory region (in bytes).
◆ unifiedAddressing
| int hipDeviceProp_t::unifiedAddressing |
Does device and host share unified address space.
◆ unifiedFunctionPointers
| int hipDeviceProp_t::unifiedFunctionPointers |
Indicates device supports unified function pointers.
◆ uuid
| hipUUID hipDeviceProp_t::uuid |
UUID of a device.
◆ warpSize
| int hipDeviceProp_t::warpSize |
Warp size.
The documentation for this struct was generated from the following file:
- include/hip/hip_runtime_api.h