Hello, I am new to GPU programming. Currently I’m trying to develop a system using CUDA C. However, my question about global memory is can different threads read the same global memory location at the same time?
Yes! You can read more about memory coalescence just by doing simple googling and that’ll teach you how global loads in CUDA work.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| read the same position in global mem | 4 | 5431 | November 6, 2007 | |
| read global memory conflict | 4 | 2751 | July 16, 2009 | |
| Single address coalescing | 2 | 9590 | January 29, 2011 | |
| global memory latency | 12 | 16419 | December 13, 2007 | |
| Moving a (BS_X+1)*(BS_Y+1) global memory matrix by BS_X*BS_Y threads | 0 | 595 | December 15, 2012 | |
| Global memory broadcasting? | 4 | 5959 | October 2, 2008 | |
| global memory access approach - not just coalescing? | 0 | 967 | March 21, 2010 | |
| memory latency | 5 | 4094 | March 21, 2007 | |
| coalesced read to shared memory | 0 | 606 | October 27, 2011 | |
| Global Memory Coalescing: Read and Write Memory Coalescing | 9 | 8383 | July 31, 2007 |