CSCI-UA.0201-003
Computer Systems Organization
Final Exam – Fall 2016 (time: 75 minutes)
Last name: First name: NetIT:
Copyright By PowCoder代写 加微信 powcoder
(Total 30 points)
1. (5 points) Circle the correct answer among the choices given. If you circle more than one answer, you will lose the grade of the corresponding question.
(A) Whenever there is a cache miss, a replacement policy must be used. a. The above statement is true.
b. The above statement is false.
c. It depends on the design of the cache.
(B) During execution, the CPU encountered a “divide by zero” event, what will happen then?
a. execution of a system call b. execution of an exception c. execution of an interrupt d. none of the above
(C) Which of the following does not decrease even if we increase cache associativity and the total cache size?
a. conflict misses
b. compulsory misses c. capacity misses
d. none of the above
(D) Getting a TLB hit means level 1 cache will not be accessed a. True
c. Depends on the OS e. Cannot tell
(E) Assume a signed int x, what does the following expression present: (1 + (x<<3) + ~x + x)
e. none of the above
2. [12 points] Given the following x86_64 assembly code and its corresponding C code, fill- in the blanks in the corresponding C code. Also on the far right, fill in the correspondence between each register and its corresponding variable in C. (Hint: you can neglect edx because it does not correspond to any variable in the C code and is used only by the compiler for the translation)
foo: xorl %eax, %eax movl $32, %ebx
movl $0, %ecx
L0: cmpl %ebx, %ecx jge L1
cmpl %edi, %eax jle L2
addl %esi, %eax jmp L3
L2: movl %edi, %edx subl %esi, %edx
addl %edx, %eax
L3: addl $2, %ecx jmp L0
int foo(int x, int y){
int j = 32 ; int sum = 0 int i;
for(i = 0 ; i < j ; i+= 2 ){ if( sum > x)
sum += y ; else
sum += x-y ; }
return sum;
edi x esi y ebx j eax sum ecx i
3. [5 points] A cache has an access time of 1 cycle. The computer with that cache experienced an average memory access time of 10 cycles, and the cache hit rate is 90%. What is the access time of the main memory? Did we benefit from having a cache in this system? Why? [A correct final answer without the steps leading to that answer will earn you a ZERO]
avg mem access = m + (1-p)M from the problem: avg = 10 m = 1 p = 0.9 M = 90 cycles
yes, we benefit from the cache because with cache avg mem access = 10 cycles and without cache it is 90 cycles
4. For the following piece of code:
void doit() {
if (fork() != 0) {
fork(); printf(“hello\n”); }
int main() { printf(“hello\n”);
doit(); printf(“hello\n”); exit(0);
a. [1 point] How many times will “Hello” be printed?
b. [1 point] Before any of the processes exit; how many stacks exist?
c. [1 point] If the system that is executing the above code is a single-core system and is using one-level page table. How many TLBs exist in the system before any of the processes execute exit(0)?
5. [5 points] Suppose we have the following C function:
void memory(){ int * p; int * q;
int * r; int * k;
p = (int *) malloc(100*sizeof(int)); 1 q = (int *) malloc(50*sizeof(int)); 2 free(p);
r = (int *) malloc(60*sizeof(int)); 3 k = (int *) malloc(50*sizeof(int)); 4
p = (int *) malloc(100*sizeof(int)); 5
….. rest of the code … }
For each one of the numbered malloc above, indicate in the following table whether that call needs to do a syscall or not. You do not need to explain. We assume that the allocator will allocate exactly what is specified in malloc and not more.
malloc # syscall Yes/No 1Y 2Y 3N 4Y 5N
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com