Even better, as user673679 and JDługosz mentioned, there is no need to have a class TimedTest, you can just write a function that takes another function as an argument and runs that in a loop. The only difference with your code would then be that you interleave running each test once, and accumulate results, and at the end calculate all the averages.

Even better, as user673679 and JDługosz mentioned, there is no need to have a class TimedTest, you can just write a function that takes another function as an argument and runs that in a loop. The only difference with your code would then be that you interleave running each test once, and accumulate results, and at the end calculate all the averages.

Avoid new an delete

Even better, as user673679 and JDługosz mentioned, there is no need to have a class TimedTest, you can just write a function that takes another function as an argument and runs that in a loop. The only difference with your code would then be that you interleave running each test once, and accumulate results, and at the end calculate all the averages.

Avoid new and delete

Even better, as user673679 and JDługosz mentioned, there is no need to have a class TimedTest, you can just write a function that takes another function as an argument and runs that in a loop. The only difference with your code would then be that you interleave running each test once, and accumulate results, and at the end calculate all the averages.

Use auto and using to avoid writing long type names

Instead of writing out full type names like in this line of code:

std::chrono::steady_clock::time_point begin = std::chrono::steady_clock::now();

I recommend you use using to create a type alias for the clock type you want to use, and auto to avoid having to specify types at all where appropriate. The above then becomes:

using clock = std::chrono::steady_clock;
auto start = clock::now();
...
auto stop = clock::now();