The Big Programming Thread - Page 336

On August 07 2013 03:03 WindWolf wrote:

http://i.imgur.com/YjuZzPx.png

This is everything I see in the properties panel (The full path is whited out by me)

On August 07 2013 03:03 WindWolf wrote:

http://i.imgur.com/YjuZzPx.png

This is everything I see in the properties panel (The full path is whited out by me)

On August 06 2013 14:01 RoyGBiv_13 wrote:


Compilers can figure out the fastest way to swap based on your architecture and language. They recognize this sort of pattern pretty well.

EDIT: confirmed: xor swaps, arithemetic swaps, an temporary variable swaps are all treated by the compiler to generate identical assembly. There is one case where this is not true, when a temporary variable is used for several uses, it is assumed the user did want to use a temporary holding state, usually a register.

On August 07 2013 04:07 spinesheath wrote:

Heh, turns out that doesn't exist for c++, I assumed it would be the same as in c# (why wouldn't it). Check out this instead:
http://stackoverflow.com/questions/1568314/copy-to-output-directory-in-vc

On August 07 2013 20:52 Morfildur wrote:
Microoptimization is the root of all evil.

int Difference(char* number1, char* number2);

return Parse(number1) - Parse(number2);

On August 08 2013 05:14 spinesheath wrote:

Since you brought it up - there's a thing that has been bugging me for a while:
Somewhere on TL, quite a while ago, someone brought up the task of implementing

int Difference(char* number1, char* number2);

I think it was part of a job interview. Now there is this trick where you subtract the characters, abusing the way ASCII sorts numbers. This seems like the intended solution for the task.

However, if I were to code this function for real, I would do it like this:

return Parse(number1) - Parse(number2);

Along with a note to the guy that requested the function that he should try to clean up his mess and parse those strings much earlier.
After all it's completely ridiculous to pass around strings as numbers but at the same time try to optimize that code.

So if I ever were to encounter such a question in a job interview or such, should I assume that they want to see if I know these cute tricks, or is it a trick question and the correct solution is what I would normally do?

On August 08 2013 06:57 CecilSunkure wrote:
Funny thing about the XOR swap, in some C++ profiling I did a while ago a simple swap function in C++:

template <typename T>
void Swap( T& a, T& b )
{
  T c = a;
  a = b;
  b = T;
}

Was actually significantly more efficient, even with integers, for some reason. The above code must have produced results the compiler knew how to optimize away. The above is also simpler and more flexible than the silly XOR trick.

On August 08 2013 07:37 RoyGBiv_13 wrote:


protip, inline this function... or it could be that the compiler already did so, and that was why it was so fast...

On August 08 2013 07:50 CecilSunkure wrote:

Well in the test we wrote the XOR inline, and then changed it out for this function. Both had no jump.

On August 08 2013 06:57 CecilSunkure wrote:
Funny thing about the XOR swap, in some C++ profiling I did a while ago a simple swap function in C++:

template <typename T>
void Swap( T& a, T& b )
{
  T c = a;
  a = b;
  b = T;
}

Was actually significantly more efficient, even with integers, for some reason. The above code must have produced results the compiler knew how to optimize away. The above is also simpler and more flexible than the silly XOR trick.

T c = a;

On August 08 2013 09:58 Nesserev wrote:

c is an object of any class T; T doesn't have to be a primitive data type.
But I don't get why the 6th line is:

b = T;

and not

b = c;