Table of Contents
Advertisement
High Level Language Optimization
for(i=0; i<5; i++){
f(i);}
is converted to the faster equivalent
f(0);
f(1);
f(2);
f(3);
f(4);
This optimization eliminates loop overhead.
Additionally, there is also a method of loop unrolling for loops with an unknown amount of
iterations at compile time. Therefore, the method is often referred to as dynamic loop unrolling. By
breaking down an arbitrary size loop into small unrolled blocks, some loop overhead can be
avoided.
For example, it is unlikely that a compiler will unroll this code.
void f(int nTotalIterations)
{
for(i=0; i<nTotalIterations; i++){
f(i);}
}
Replacing this small loop with the considerably larger code segment below will potentially provide
a significant performance improvement at the expense of code size.
void f(int nTotalIterations)
{
const int nItersPerBlock = 4;
int nTotalBlockIters;
int i;
// find the largest multiple of nItersPerBlock that is less
//
nTotalBlockIters = (nTotalIterations / nItersPerBlock) *
5-10
than or equal to nTotalIterations
Intel® PXA27x Processor Family Optimization Guide
Advertisement
Table of Contents
