vector function 'at'

Since the other direct access methods front(), back() and []开发者_C百科 do not check for bounds, does 'at' introduce too much overhead for these checking if the vectors are large and one wants to iterate over them sequentially? I guess if one is sure of the bounds, [] is a better direct access option.

That's why there're separate [] and at(). If you're sure that you don't access out of bounds you use [] (or iterators), otherwise you may use at() if you wish.

at() cannot be quicker than [], but "too much overhead" is really up to you to decide. I suggest profiling the code with [] and at() and see whether it makes a significant difference.

If you want to iterate sequentially, then do neither.

It is preferable to use:

// C++03
BOOST_FOREACH(T const& t, vec) { }

// C++0x
for (T const& t: vec) { }

You'll get the maximum speed with no risk of stepping outside the defined range.

at method is slightly slower than the subscript [] operator. Here is a simple timing test.

#include <ctime>
#include <cstdio>
#include <vector>
#include <stdint.h>
#include <algorithm>
using namespace std;

// Pentium clock timer
extern "C" {
  __inline__ uint64_t rdtsc() {
    uint32_t lo, hi;
    __asm__ __volatile__ (
    "xorl %%eax,%%eax \n        cpuid"
    ::: "%rax", "%rbx", "%rcx", "%rdx");
    __asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi));
    return (uint64_t)hi << 32 | lo;
  }
}

vector<int>A;

uint64_t time_at()
{
    uint64_t start=rdtsc();
    srand(11101);
    for(int i=0;i<10000000;i++)
        int x=A.at(rand()%10000000);
    uint64_t end=rdtsc();
    return end-start;
}

uint64_t time_subscript()
{
    uint64_t start=rdtsc();
    srand(11101);
    for(int i=0;i<10000000;i++)
        int x=A[(rand()%10000000)];
    uint64_t end=rdtsc();
    return end-start;
}

int main()
{
    for(int i=0;i<10000000;i++)
        A.push_back(i);

    printf("At ===> %llu\n",time_at());
    printf("[] ===> %llu\n",time_subscript());
    return 0;
}

For one, at throws exception if the index is out of bounds, therefore requiring a compilation with exceptions enabled. That's likely one of the reasons it is just an additonal method for accessing vector elements.