Using CUDA Occupancy Calculator -
using occupancy calculator cannot understand how registers per thread / shared memory per block .i read documentation.i use visual studio .so in project properties under cuda build rule->command line -> additional options add --ptxas-options=-v.the program compiles fine .but not see output .can help?
with switch on there should line on compiler output window tells number of registers , amount of shared memory.
see @ on compiler output window? can copy , paste question?
should
ptxas info : used 3 registers, 2084+1060 bytes smem, 40 bytes cmem[0], 12 bytes cmem[1]
Comments
Post a Comment