HW #4. Processes, Threads & Race Conditions Assignment (60 points). For this assignment you will create a group of threads and then try to coax them into producing as many race conditions as possible. First, your program should declare a single shared integer variable (i.e. a global variable) with initial value zero; then your program should spawn off the number of threads specified in the source file using the MAXTHREADS defined constant. We will use a for-loop to create these threads. We will designate threads created on even numbered iterations as adders and threads created on odd iterations as subtractors. DO NOT CHANGE THE VALUE OF THE CONSTANCE DOING SO WILL CUASE A 15 POINT PENALTY. Adder Threads: (Adding threads WILL have their own thread function) Adder threads will execute the following steps: 1. Read global value and store it into temp 2. add 3 to the temp value 3. store temp value into global variable 4. print the following to the console: Current Value written to Global Variables by ADDER thread id: tid is val Were tid is the thread id and val is the value of the temporary variable AFTER adding 3 to it. 5. repeats steps 1 through 4 for MAXITERATIONS (this is a defined constant in the given source code). Subtractor Threads: (Subtracting threads WILL have their own thread function) Subtractor threads will execute the following steps: 1. Read global value and store it into temp 2. subtract 3 from the temp value 3. store temp value into global variable 4. print the following to the console: Current Value written to Global Variables by SUBTRACTOR thread id: tid is val Were tid is the thread id and val is the value of the temporary variable AFTER subtracting 3 from it. 5. repeats steps 1 through 4 for MAXITERATIONS (this is a defined constant in the given source code). Once all threads have finished, please print the value in main in the following way: Final Value of Shared Variable : val Where val is the value stored in the shared variable. Make sure to wait for all threads to finish and not just one. Since these threads are unsynchronized, there is a possibility of race conditions occurring depending on the actual interleaved execution order of the various threads. Insert nanosleep() commands into your thread code to induce as many race conditions as possible by forcing an (in)appropriate interleaved execution sequence. A timespec struct has been given to you. You can change the values in the struct. The first value is the amount of time to sleep in seconds and the second value is the amount of time to sleep in nanoseconds. Modify these values to try to get the final value the furthest away from 0 zero as possible in either the negative or positive direction. As your program executes, it will produce a trace of output from each thread showing the actual interleaved order of execution Where you place the nanosleep() call will affect the amount of race conditions that occur and how much your print statements are interleaved. Implement your solution using Pthreads on Linux

Here is part of the program:

#include #include #include #include #include #include /** * THESE DEFINE VALUES CANNOT BE CHANGED. * DOING SO WILL CAUSE POINTS TO BE DEDUCTED * FROM YOUR GRADE (15 points) */ /** BEGIN VALUES THAT CANNOT BE CHANGED */ #define MAX_THREADS 16 #define MAX_ITERATIONS 40 /** END VALUES THAT CANNOT BE CHANGED */

/** * use this struct as a parameter for the function * nanosleep. * For exmaple : nanosleep(&ts, NULL); */ struct timespec ts = {0, 123456};

int main(int argc, char** argv) {

return 0; }