cudaDeviceProp

asyncEngineCount canMapHostMemory clockRate computeMode concurrentKernels deviceOverlap ECCEnabled integrated kernelExecTimeoutEnabled l2CacheSize major maxGridSize maxTexture1D maxTexture1DLayered maxTexture2D maxTexture2DLayered maxTexture3D maxThreadsDim maxThreadsPerBlock maxThreadsPerMultiProcessor memoryBusWidth memoryClockRate memPitch minor multiProcessorCount name pciBusID pciDeviceID pciDomainID regsPerBlock sharedMemPerBlock surfaceAlignment tccDriver textureAlignment totalConstMem totalGlobalMem unifiedAddressing warpSize

int cudaDeviceProp::warpSize Warp size in threads