This slide explains why we want to scale down the problem effectively by giving assignment 3 as an example where we have only limited number times to test our program on Blacklight due to long queue. Because of this limitation, we want to debug/test and make sure our code runs correctly in a small CPU machine before testing its performance in a bigger machine. Moreover, because it takes huge computational cost to diagnose the program as we experienced in assignment 2 where using printf costed severely to the performance, we do not want to debug and test on a full data set, but rather on a subset of data.
This slide explains why we want to scale down the problem effectively by giving assignment 3 as an example where we have only limited number times to test our program on Blacklight due to long queue. Because of this limitation, we want to debug/test and make sure our code runs correctly in a small CPU machine before testing its performance in a bigger machine. Moreover, because it takes huge computational cost to diagnose the program as we experienced in assignment 2 where using printf costed severely to the performance, we do not want to debug and test on a full data set, but rather on a subset of data.
This comment was marked helpful 0 times.