fix for python3 int division tp_reverse.py

CUDA/tp_reverse.py

@@ -24,7 +24,7 @@ size_data=100
 # number of threads in a block
 block_size=32
 # number of blocks of thread
-num_blocks=(size_data+block_size-1)/block_size
+num_blocks=(size_data+block_size-1)//block_size
 
 source = numpy.arange(0,size_data,1).astype(numpy.float32)
 destination = numpy.zeros_like(source)