Hello, I am trying to simulate a multiprocessor system with distributed shared memory (NUMA). To get started, I want to know the following things : 1. Which of the existing cache coherence protocols do I use ? I think MOESI_hammer, but not sure. Or do I need to write a separate protocol ? 2. Also, can I reuse any of the sample configuration scripts (se.py with --ruby switch) ? I don't intend to simulate a full system, but an Alpha system in syscall emulation mode only. I have written a few basic scripts while learning, but haven't figured out how to add different memories along with full-bit directory to each cpu. I came across a patch that enables distributed memory by configuring individual memories as SimpleMemory and a PhysicalMemory to encapsulate all these memories, but haven't seen any config script making use of this.
-- Regards; Abhishek Bhowmick
_______________________________________________ gem5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
