; 120  :         volatile __int64 x;
; 121  : 
; 122  :         x = 1;

  00037    c7 45 e0 01 00
    00 00         mov     DWORD PTR _x$155193[ebp], 1
  0003e    89 75 e4     mov     DWORD PTR _x$155193[ebp+4], esi

volatile long flag;

flag = 1;

long f = flag;

flag ++;

flag ++;

++flag ;

Objects declared as volatile are not used in certain optimizations because their values can change at any time. The system always reads the current value of a volatile object at the point it is requested, even if a previous instruction asked for a value from the same object. Also, the value of the object is written immediately on assignment. 

Also, when optimizing, the compiler must maintain ordering among references to volatile objects as well as references to other global objects. In particular, 

A write to a volatile object (volatile write) has Release semantics; a reference to a global or static object that occurs before a write to a volatile object in the instruction sequence will occur before that volatile write in the compiled binary.

A read of a volatile object (volatile read) has Acquire semantics; a reference to a global or static object that occurs after a read of volatile memory in the instruction sequence will occur after that volatile read in the compiled binary.

This allows volatile objects to be used for memory locks and releases in multithreaded applications.

volatile __int64 x;
.............
x = 0i64;

flag++;

++flag;

В VS2005 volatile дополнительно к подавлению оптимизации работает как барьер соответствующей операции для самого компилятора (но не для процессора).
Никакой атомарности для произвольных типов он не обеспечивает.

long oldValue = flag;
flag = newValue;
return oldValue;

while (global_flag){ ... }

Example 8-1. Stores Are Not Reordered with Other Stores
---------------------------------
Processor 0 | Processor 1
---------------------------------
mov [ _x], 1 | mov r1, [ _y]
mov [ _y], 1 | mov r2, [ _x]
---------------------------------
Initially x == y == 0
r1 == 1 and r2 == 0 is not allowed
---------------------------------

#define MAX_THREAD 4
#define MAX_INC 10000000

/* bool == BYTE == char */
long a = 0; /*char a = 0; */
volatile long b = 0; /*volatile char b = 0;*/
long c = 0;

DWORD WINAPI start(LPVOID lpThreadParameter)
{
    int i = 0;
    for (i = 0; i < MAX_INC; i++)
    {
        a++;
        b++;
        InterlockedIncrement(&c);
    }
}

int main(int argc, char *argv[])
{
    HANDLE h[MAX_THREAD];
    DWORD n = 0;
    int i = 0;

    for (i = 0; i < MAX_THREAD; i++)  {
        h[i] = CreateThread(NULL, 0, start, NULL, 0, &n);
        if ( INVALID_HANDLE_VALUE == h[i]) ExitProcess(-1);
    }
    WaitForMultipleObjects(MAX_THREAD, &h, TRUE, INFINITE);
    printf("%i \t a = %i \t b = %i \t c = %i", MAX_THREAD*MAX_INC, a, b, c);
    /*printf("%i \t a = %i \t b = %i \t c = %i", 0xff & (MAX_THREAD*MAX_INC), a, b, (0xff & c) ); */

    getchar();
    return 0;
}