??? 12/07/02 14:25 Read: times |
#33993 - RE: Multiprocessor Communications |
Eeeek!!! multiprocessor communications...
some pitfalls are: Deadlock situations where a process on one processor is waiting for another on another processor to finish,and the second is waiting for the first to finish. Checking shared resources with semaphores. and or spinlocks. Accesing shared memory with semaphores. task scheduling and task allocation. all these and other fun things make multiprocessor system design highly specialised I worked on a system which used 8 strongarms on a PCI board,which is fine till you try to deside things like which is the master,how you tell when a processor is idle,when a processor can access shared memory safely.Etc Good luck,you'll need it. Jez |
Topic | Author | Date |
Multiprocessor Communications | 01/01/70 00:00 | |
RE: Multiprocessor Communications | 01/01/70 00:00 | |
RE: Multiprocessor Communications | 01/01/70 00:00 | |
RE: Multiprocessor Communications![]() | 01/01/70 00:00 | |
RE: Multiprocessor Communications | 01/01/70 00:00 | |
RE: Multiprocessor Communications | 01/01/70 00:00 | |
RE: Multiprocessor Communications | 01/01/70 00:00 | |
RE: Multiprocessor Communications | 01/01/70 00:00 | |
RE: \\ | 01/01/70 00:00 | |
RE: \\ | 01/01/70 00:00 | |
RE: Multiprocessor Communications | 01/01/70 00:00 | |
RE: Multiprocessor Communications | 01/01/70 00:00 |