Empirical investigation of cloud, grid and virtualization using compiler optimization level for threads processes

This research focused on implementation of OpenMP. It considers the parallelization of an application code which simulates the thermal gradient of a material in two dimensions. A C language program code called jacobi2d.c that solves a rectangular 2 dimensional heat conductivity problem using Jacobi iterative method was used. The boundary conditions required to compute a temperature distribution for a rectangular 2D problem are: Top 300C, Bottom 500C, Left 400C and Right 900C with a range of problem sizes enter as a run-time parameters to alter the problem sizes and convergence criteria. Also, there were computations and readings for iterations and runtime for four values of M and N which was selected for 01, 02 and 03 optimizations. In Table 1.1 Readings, four values were selected for each of the iterations. The results shows the performance of the runtime as the processor increases from 01 -optimization, to 02 optimization and finally to 03-optimization. It can be deduce from the representation that the run time of the values reduces as more resources are allocated to execution through the increase in optimization level. Also, in Table 2.1 Readings, the runtime decreases as it moves from thread1, thread2, thread3 and thread4, comparing the last values for thread1 which are M is 180,N is 200, and their runtime which is 42.797187001 . Also the last values for thread2 which are M is 180, N is 200, their runtime which is 21.772106003. When the two runtimes were compared, it was discovered that there was a decrease in the runtime because the more the thread increases, the more system resources they share such as a processor which may affect their runtime by increasing it.