Use shared Memory

Accelerated Computing CUDA CUDA Programming and Performance
1

Hi,

I am confused about how to use shared memory.
As I know shared memory is a memory space shared by all the threads of the same blocks, and when When we declare an array as shared the compiler creates a copy for each thread of the launched block.

I need to know if it is possible that each thread may have a different array content in the shared memory ?

Thanks in advance

2

this may be of interest:

https://stackoverflow.com/questions/58380766/is-it-possible-to-dedicate-a-portion-of-shared-memory-to-each-thread

3

Thank you

4

I am confused about how to use shared memory.
I have found this article to be very good: Using Shared Memory in CUDA C/C++ | NVIDIA Technical Blog