Hello everyone, I’m a rookie and have some questions about sharedMemPerBlock
My GPU is GeForce GTX 750 Ti, Cuda v9.2
I call the function cudaGetDeviceProperties , get the value of struct cudaDeviceProp, in which the member variable sharedMemPerBlock is equal to 49152, variable type is unsigned__int64, so I think the size is 49152 * 8 = 393216 bytes
but I try to use the Block-Shared-Memory, it can’t be bigger than 49152 bytes ,I am confused here
can someone help me? thanks a lot:)