OSdevelopment
Development related to Operating Sytems algorithms and techniques:
- Threads
- Barriers
- Parallel programming synchronization
- Process Scheduling
- Memory Management
1. CPU-schedulers
This app consists of two parts:
An implementation of a shell (ep1sh.c), using the fork() syscall and a loop structure. It has the commands date and chown : built-in (using exactly the structure provided before). It also executes any command using the execvp() syscall.
And also, a multithreaded implementation of three process scheduling algorithms for CPU's, namely the Shortest Job First, Round Robin and Priority Scheduler. The program will simulate a list of processes providaded in a trace file via command line arguments. It also supports a debug flag 'd', to print the events as they occur.
2. multithreadedSync
This is an implementation of a velodrome bike racing simulator, fully concurrent (each biker is a thread), and all the threads are synchronized using barriers, in our case, using the pthread_barrier from the pthread library. The race is simulated with strict rules for the scores of each biker, for details of the rules, please refer to the assigment.pdf file (or e-mail me :p). Each biker has some possible moves, and each one of them behave like an automata, and we synchronize the movements of the bikers using mutexes, namely pthread_mutexes.
The simulator constructs a dependency graph each step, and calculates the Strongly Connected Components to detect cycles in the graph using Tarjan's Algorithm, to prevent deadlocks in the simulation.
3. memory-simulator
Simulator of virtual and physical memory allocation, using a trace file for the processes list (each process has a default size and make some memory access in a given time). The memory-simulator implements the following algorithms:
Virtual memory allocation:
- Best Fit
- Worst Fit
- Quick Fit
Pages substitution(for page faults):
- Optimal
- FIFO
- LRU (bit matrix implementation)
- LRU (bit counter implementation)
