Measuring exception handling overhead in C++

Question

What is the best way to measure exception handling overhead/performance in C++?

Please give standalone code samples.

I\'m targeting Microsoft Visual C++ 2008

Answer 1

Full details on how g++ handles exceptions is shown here. It describes it as being for the Itanium architecture, however the general techniques used are the same. It won't tell you exact overhead in terms of time, however you can glean what the rough code overhead will be.

Answer 2

There's no really good way to meassure that in code. You would need to use a a profiler.

This won't show you directly how much time is spent with exception handling but with a little bit of research you will find out which of the runtime methods deal with exceptions (for example for VC++.NET it's __cxx_exc[...]).

Add their times up and you have the overhead. In our project we used vTunes from Intel which works with both Visual C++ and gcc.

Edit: Well, if you just need a generic number that might work. Thought you had an actual application to profile where you can't just turn off exceptions.

Answer 3

Another note on exception handling performance: simple tests don't take caching into account. The try-code and the catch-code are both so small that everything fits in the instruction and data caches. But compilers may try to move the catch-code far away from the try-code, which reduces the amount of code to keep in cache normally, thus enhancing performance.

If you compare exception handling to traditional C-style return-value checking, this caching effect should be taken into account as well (the question is usually ignored in discussions).

Carl

Answer 4

As a suggestion: don't bother too much with the overhead when exceptions are thrown. Exception handling implementations usually make not throwing fast and catching slow. That's ok since those cases are, well, exceptional.

Carl

Answer 5

Kevin Frei talks about exception handling performance cost in his speech "The Cost of C++ Exception Handling on Windows". (Under "Summary & Conclusions" there is one list item that says "[Exception handling performance cost is] not always measurable".)

Answer 6

Here's the measuring code I came up with. Do you see any problems with it?

Works on Linux and Windows so far, compile with:

g++ exception_handling.cpp -o exception_handling [ -O2 ]

or for example Visual C++ Express.

To get the base case ("exception support removed from the language altogether"), use:

g++ exception_handling.cpp -o exception_handling [ -O2 ] -fno-exceptions -DNO_EXCEPTIONS

or similar settings in MSVC.

Some preliminary results here. They are probably all hokey because of varying machine load, but they do give some idea about the relative exception handling overhead. (Executive summary: none or little when no exceptions are thrown, huge when they actually are thrown.)

#include <stdio.h>

// Timer code

#if defined(__linux__)
#include <sys/time.h>
#include <time.h>

double time()
{
    timeval tv;
    gettimeofday(&tv, 0);
    return 1.0 * tv.tv_sec + 0.000001 * tv.tv_usec;
}
#elif defined(_WIN32)
#include <windows.h>

double get_performance_frequency()
{
    unsigned _int64 frequency;
    QueryPerformanceFrequency((LARGE_INTEGER*) &frequency); // just assume it works
    return double(frequency);
}

double performance_frequency = get_performance_frequency();

double time()
{
    unsigned _int64 counter;
    QueryPerformanceCounter((LARGE_INTEGER*) &counter);
    return double(counter) / performance_frequency;
}
#else
# error time() not implemented for your platform
#endif

// How many times to repeat the whole test
const int repeats = 10;

// How many times to iterate one case
const int times = 1000000;

// Trick optimizer to not remove code
int result = 0;



// Case 1. No exception thrown nor handled.

void do_something()
{
    ++result;
}

void case1()
{
    do_something();
}



// Case 2. No exception thrown, but handler installed

#ifndef NO_EXCEPTIONS
void do_something_else()
{
    --result;
}

void case2()
{
    try
    {
        do_something();
    }
    catch (int exception)
    {
        do_something_else();
    }
}



// Case 3. Exception thrown and caught

void do_something_and_throw()
{
    throw ++result;
}

void case3()
{
    try
    {
        do_something_and_throw();
    }
    catch (int exception)
    {
        result = exception;
    }
}
#endif // !NO_EXCEPTIONS

void (*tests[])() =
{
    case1,
#ifndef NO_EXCEPTIONS
    case2,
    case3
#endif // !NO_EXCEPTIONS
};

int main()
{
#ifdef NO_EXCEPTIONS
    printf("case0\n");
#else
    printf("case1\tcase2\tcase3\n");
#endif
    for (int repeat = 0; repeat < repeats; ++repeat)
    {
        for (int test = 0; test < sizeof(tests)/sizeof(tests[0]); ++test)
        {
            double start = time();

            for (int i = 0; i < times; ++i)
                tests[test]();

            double end = time();

            printf("%f\t", (end - start) * 1000000.0 / times);
        }
        printf("\n");
    }

    return result; // optimizer is happy - we produce a result
}