1) Perform a full system monitoring process using the tools that you think are most adequate and reach a diagnostic on the use of resources and the bottleneck that might exist in the system. Simulate the system's workload of the code of sumdis.c given in the unit that covers the clusters. For example, use:
sumdis 1 2000000
2) Change the parameters of the kernel and the compiler and execute the code mentioned in the preceding point (sumdis.c) with, for example:
time ./sumdis 1 1000000
3) The same with both kernels and formulate a conclusion regarding the results.