In this project a chemical reaction takes places which is called fusion reaction takes places in sun in which number of hydrogen atoms combine with each other to make deuterium, helium3 and helium4 atoms. We provide a number of hydrogen atoms in a specific sized box and calculate the time in which almost all hydrogen atoms converting in to Helium. When we Parallelize this fusions reaction code then time may be reducing effectively.
Fusion reaction is a type of chemical reaction in which hydrogen atoms are colluding themselves and releasing very large amount of energy. This reaction requires a very huge amount of temperature to take place (>=5000) which is difficult to achieve in our daily life and in our liberties, therefore we write a computer program which simulates this reaction according to the requirement and shows all the possible events that happens during introduction.
Fusion simulation project is firstly sequentially prepared by the team itself and after thoroughly study they start implementing it and by the end of due that they submit their working project.
Implementation
Hydrogen atoms posses random motion and each atom can collide with any other atom at any time, to keep track of each and every single atom was a challenging task but team worked hard and did it before mid project presentation. But it is very slow and take alot of time when the data size get increases.
For example each atom have its own coordinates, velocity and these all parameters are variable and changing with every instant of time.
To detect collision between different atoms Pythagoras theorem is used. The methodology was that we compute that if the distance which is calculating through this theorem is less than the diameter of the balls then it means a collision has occur.
Parallel Implementation
Accurate calculation of temperature in sequential mode takes a lot of time which is not desired by engineers. Therefore we designed a parallel algorithm and implemented it to reduce the time taken for temperature calculation.
Algorithm
Firstly we will have this grid shown below for parallelism we divide the whole data into all available nodes.
Architecture for parallel program is little bit complex, firstly total data size (hydrogen atoms) are divided into all available node and then if there are further cores in the node then on each core a thread will run for complete utilization.
Challenges
1. Development of sequential code was a big challenge it is because we did not do any type of simulation before.
2.Communication between different processes. It was an overhead in performance to solve this issue we use various algorithms and finally got the solution.
Output:
The final output of the program shows a frame in which thousands of hydrogen atoms are moving colliding among themselves and creating new atoms.
Speedup Graph
Execution Time Graph
