Return CUDA device info

dinfo = cv.Utils.deviceInfo(device)

Input

• device System index of a CUDA device starting with 0.

Output

• dinfo scalar struct of information about CUDA device:
  • deviceID system index of the CUDA device starting with 0
  • name ASCII string identifying device
  • totalGlobalMem global memory available on device in bytes
  • sharedMemPerBlock shared memory available per block in bytes
  • regsPerBlock 32-bit registers available per block
  • warpSize warp size in threads
  • memPitch maximum pitch in bytes allowed by memory copies
  • maxThreadsPerBlock maximum number of threads per block
  • maxThreadsDim maximum size of each dimension of a block
  • maxGridSize maximum size of each dimension of a grid
  • clockRate clock frequency in kilohertz
  • totalConstMem constant memory available on device in bytes
  • majorVersion major compute capability
  • minorVersion minor compute capability
  • textureAlignment alignment requirement for textures
  • texturePitchAlignment pitch alignment requirement for texture references bound to pitched memory
  • multiProcessorCount number of multiprocessors on device
  • kernelExecTimeoutEnabled specified whether there is a run time limit on kernels
  • integrated device is integrated as opposed to discrete
  • canMapHostMemory device can map host memory with cudaHostAlloc/cudaHostGetDevicePointer
  • computeMode compute mode, one of:
    • Default default compute mode (Multiple threads can use cudaSetDevice with this device)
    • Exclusive compute-exclusive-thread mode (Only one thread in one process will be able to use cudaSetDevice with this device)
    • Prohibited compute-prohibited mode (No threads can use cudaSetDevice with this device)
    • ExclusiveProcess compute-exclusive-process mode (Many threads in one process will be able to use cudaSetDevice with this device)
  • maxTexture1D maximum 1D texture size
  • maxTexture1DMipmap maximum 1D mipmapped texture size
  • maxTexture1DLinear maximum size for 1D textures bound to linear memory
  • maxTexture2D maximum 2D texture dimensions
  • maxTexture2DMipmap maximum 2D mipmapped texture dimensions
  • maxTexture2DLinear maximum dimensions (width, height, pitch) for 2D textures bound to pitched memory
  • maxTexture2DGather maximum 2D texture dimensions if texture gather operations have to be performed
  • maxTexture3D maximum 3D texture dimensions
  • maxTextureCubemap maximum Cubemap texture dimensions
  • maxTexture1DLayered maximum 1D layered texture dimensions
  • maxTexture2DLayered maximum 2D layered texture dimensions
  • maxTextureCubemapLayered maximum Cubemap layered texture dimensions
  • maxSurface1D maximum 1D surface size
  • maxSurface2D maximum 2D surface dimensions
  • maxSurface3D maximum 3D surface dimensions
  • maxSurface1DLayered maximum 1D layered surface dimensions
  • maxSurface2DLayered maximum 2D layered surface dimensions
  • maxSurfaceCubemap maximum Cubemap surface dimensions
  • maxSurfaceCubemapLayered maximum Cubemap layered surface dimensions
  • surfaceAlignment alignment requirements for surfaces
  • concurrentKernels device can possibly execute multiple kernels concurrently
  • ECCEnabled device has ECC support enabled
  • pciBusID PCI bus ID of the device
  • pciDeviceID PCI device ID of the device
  • pciDomainID PCI domain ID of the device
  • tccDriver true if device is a Tesla device using TCC driver, false otherwise
  • asyncEngineCount number of asynchronous engines
  • unifiedAddressing device shares a unified address space with the host
  • memoryClockRate peak memory clock frequency in kilohertz
  • memoryBusWidth global memory bus width in bits
  • l2CacheSize size of L2 cache in bytes
  • maxThreadsPerMultiProcessor maximum resident threads per multiprocessor
  • freeMemory gets free memory
  • totalMemory gets total device memory
  • supports Struct which provides information on CUDA feature support. Is true if the device has the specified CUDA feature. Otherwise, it is false:
    • Compute10 Compute Capability 1.0
    • Compute11 Compute Capability 1.1
    • Compute12 Compute Capability 1.2
    • Compute13 Compute Capability 1.3
    • Compute20 Compute Capability 2.0
    • Compute21 Compute Capability 2.1
    • Compute30 Compute Capability 3.0
    • Compute32 Compute Capability 3.2
    • Compute35 Compute Capability 3.5
    • Compute50 Compute Capability 5.0
    • GlobalAtomics same as 'Compute11'
    • SharedAtomics same as 'Compute12'
    • NativeDouble same as 'Compute13'
    • WarpShuffleFunctions same as 'Compute30'
    • DynamicParallelism same as 'Compute35'
  • isCompatible Checks the CUDA module and device compatibility. Is true if the CUDA module can be run on the specified device. Otherwise, it is false.