Description:

In this assignment I wrote my own distributed shared memory implementation. The basic idea is that you create (through software) the illusion that two processes on different machines share the same address space. So when a process writes to a memory location in a page that resides on a remote machine, my implementation asks my other implementation on the remote machine for a copy, the remote machine should protect its page, and then the local machine maps the page's data into its own address space. If the remote machine were to access to the page, the same protocol transfers the page back.

My implementation implements the process consistency memory model (x86 memory consistency model --- (writes from one processor need to be seen in the same order by the other machine).

Thats good for description, but how to run?

For linux platforms:

To run a test case type ./test master masterip slaveip testcase_number on the master machine and ./test slave masterip slaveip testcase_number on the slave machine where masterip == ip address of master machine slaveip == ip address of slave machine testcase_number is between 0 and 5.