SE250:lab-3:stsa027

1.

The code I typed was transferred to the 'arraylist.c' the lecturer wrote for us, since compiling in my own *.c file gave some mysterious error.

This was the code used to perform task 1:

int main (){
    ArrayList ar;
    clock_t x, y;
    element_t elt = 1;
    int i;

    arraylist_init (&ar);
    x = clock();
    for (i = 0; i < 5000000; i++){
	arraylist_push (&ar, elt);
    }

    y = clock() - x;

    printf("time taken = %d", y);

    return 0;
}

The result returned when 5 million repeats were used was:

time taken = 249

The result returned when 10 million repeats were used was:

time taken = 452

The result returned when 20 million repeats were used was:

time taken = 827

As the loop number increased, the time taken also increased as expected. Also, the increase in time is roughly proportional to the number of loops - each doubling of repeats accompanies an almost doubling of time taken.

2.

When the factor was changed to 1.1,

time taken = 467

When the factor was changed to 1.5,

time taken = 249

When the factor was changed to 3.0,

time taken = 187

When the factor was changed to 3.0 up to 20.0,

time taken = 187

When the factor was changed to 70.0,

time taken = 156

I tried with values up to 2000.0, and still time taken hovered around 156ms for 5 million repeats. I think this is because as soon as the capacity is expanded, the array already has enough memory for 5 million values, which doesn't require another increase in array capcity, which may then deplete available memory and return an error.

3.

when ARRAYLIST_MIN_ALLOC is 1,

time taken = 203

when ARRAYLIST_MIN_ALLOC is 2,

time taken = 218

when ARRAYLIST_MIN_ALLOC is 4,

time taken = 202

when ARRAYLIST_MIN_ALLOC is 8,

time taken = 218

when ARRAYLIST_MIN_ALLOC is 32,

time taken = 218

As shown, changing ARRAYLIST_MIN_ALLOC size does not seem to significantly affect the time taken. I think this is because the function arraylist_put doubles the array size each time, which means that to increase the array size to 16, the ensure_capacity function needs only run 8 times, which presumably takes very little time compared to running the entire for loop which has the preset array size of 16.

4.

when I made the size of the array 5 million before running the for loop,

time taken = 156

This prevents having to increase capacity of the array during running of the for loop, which slightly decreases the run time required by about 20ms.

5.

The number of loops remains at 5 million with array capacity growing at 1,000 each time:

time taken = 4772

This is expected, because with the array growing at a constant value, the array capacity must be used many many more times (~ no. of repeats / incremental value) which involves reallocating memory, can calling the ensure_capacity function and increases time required to run the program.

6.

In order for the program to complete in a reasonable time, the number of repeats had to be reduced to 100,000 (by a factor of 50). The new values were inserted at the beginning of the array, at index 0 as instructed.

time taken = 3744

Despite the reduction in loop number, the program still took seconds to run. This is most likely due to the repeated memmove required in order to insert the new member into the beginning of the array.

Reflective Bit

I still require some more time for me to be familiar with the syntax of C++ and writing code that is more easily read. (The tutor recommended that I get used to indenting and putting the brackets in the right place.) Luckily, this lab doesn't require too much code to be written ourselves, which helped me complete the lab in a reasonable time. The fact that there were clear instructions for us to change various parameters was also good, in that it got me through the lab much more quickly, and still observed the impact that various changes made.