I was reading new materials ahead I came to know the term "Philosophers Synchronization Algorithm", but I could not understand it. Can anyone help me understand it what is it?
Thanks
+1
A:
It's just one of the many examples used to describe what can happen in a concurrent world in which you have many entities that can perform actions on shared objects without caring about each other.
The problem is simple: you have X philosophers arranged in a round table (with a fictional spaghetti dish each to be eaten) and X forks, one between every pair of philosophers.
The rules of the game impose that a philosopher needs two forks to be able to consume his spaghetti and the example shows how simply allowing any of them to try to eat without caring about anyone else can lead to
- deadlocks: every philosopher takes his left fork and then they all wait for another fork but no selfish philosopher will drop his one, so they're gonna wait forever
- starvation: there's no guarantee that any philosopher will eventually be able to eat (check wikipedia page for exact explaination of why)
- livelocks: another classical example.. if a rule imposes to phils to try to get a second fork after getting the first one for max 5 minutes, then release the already acquired one you can have a situation in which all of them are exactly synched and they keep taking one fork and releasing it after time expires
In you question you clearly speak about an algorithm related to this problem (so I suppose an algorithm meant to solve the just described problems), wikipedia offers 4 of them here.
Jack
2010-07-13 22:03:47