Given C code: void psum1(float a[], float p[], long n) { long i; p[0] = a[0]; for(i = 1; i<n ;i++); p[i] = p[i-1]+a[i]; } Cycles per element (CPE): The CPE denotes performance of program that helps in improving code. It helps to understand detailed level loop performance for an iterative program. It is appropriate for programs that use a repetitive computation. The processor’s activity sequencing is controlled by a clock that provides regular signal of some frequency. Memory aliasing: It denotes a case where two pointers might designate to memory location that is same. The compiler should assume that unlike pointers might be aliased for safe optimizations. The program aspects that could limit chances for a compiler in generating optimized code denote optimization blockers. If a compiler is unable to determine whether two pointers might be aliased, it should adopt that either case would be possible. The possible optimization set is been limited in this case.

Question

Want to see the full answer?

Answer 1

Question

Chapter 5.12, Problem 5.12PP

Program Plan Intro

Given C code:

void psum1(float a[], float p[], long n)

{

long i;

p[0] = a[0];

for(i = 1; i<n ;i++);

p[i] = p[i-1]+a[i];

}

Cycles per element (CPE):

The CPE denotes performance of program that helps in improving code.

It helps to understand detailed level loop performance for an iterative program.

It is appropriate for programs that use a repetitive computation.

The processor’s activity sequencing is controlled by a clock that provides regular signal of some frequency.

Memory aliasing:

It denotes a case where two pointers might designate to memory location that is same.

The compiler should assume that unlike pointers might be aliased for safe optimizations.

The program aspects that could limit chances for a compiler in generating optimized code denote optimization blockers.

If a compiler is unable to determine whether two pointers might be aliased, it should adopt that either case would be possible.

The possible optimization set is been limited in this case.

Expert Solution & Answer

Want to see the full answer?

Check out a sample textbook solution

See solution

Students have asked these similar questions

12. (identities). Minimize the function F using algebraic modifications (w,x,y,z) Y(Z+WX (W+Z)) + (XY+(x+y) + Z) + z ((ZX +2X) A M+1) = 12²x²x²m)_]

Chapter 3: Practice Problems VIII. You are given the following C code to compute integer absolute value: int abs(int x) { } return x > (sizeof(int)*8-1); int comp = x mask; return comp; } A. What bit pattern does mask have, as a function of x? B. What numerical value does mask have, as a function of x? C. For what values of x do functions abs and opt abs return identical results? D. For the cases where they produce different results, how are the two results related? E. Show that with the addition of just one single arithmetic operation (any C operation is allowed) that you can fix opt abs. Show your modifications on the original code. (You can just provide the line that you will add). F. Are there any values of x such that abs returns a value that is not greater than 0? Which value(s)?

(4) [8]Write code for a function multBy3divBy4, that for argument int x, computes 3*x/4, but follows the bit-level integer coding rules (text p.128) like in problem 3 of the assignment 1. (you may assume ints are 32 bits). (Note: The operation in your code 3*x is allowed to cause overflow.)