Код: Выделить всё
/*
* Compile on 64-bit Linux or Solaris (I guess it will probably also work on
* FreeBSD), like this:
* g++ -o bin2dec bin2dec.cpp -std=c++11 #Don't put -O3
* here, as GCC seems to ignore `volatile` in global variables.
* Or like this:
* clang++ -o bin2dec bin2dec.cpp -O3
* Playing with inline assembly is a lot easier on UNIX-like systems than on
* Windows.
* */
#include
#include
extern "C" {
char binary_input[9]; // I think this does not need to be volatile because the
// inline assembly isn't changing it. It's only reading
// from it.
volatile int first_digit,
binary_coded_decimal; // GCC (but not CLANG) clobbers the `first_digit` if
// you turn on the optimizer in spite of the
// `volatile`.
}
int main() {
std::cout > binary_input;
asm(R"assembly(
.intel_syntax noprefix
mov r8, 0
mov r9, 0
input_loop:
cmp byte ptr [binary_input + r9], '0'
jz input_is_zero
cmp byte ptr [binary_input + r9], '1'
jz input_is_one
jmp end_of_the_input_loop # The current character is presumably '\0'.
input_is_one:
shl r8, 1
inc r8
jmp end_of_branching
input_is_zero:
shl r8, 1
jmp end_of_branching
end_of_branching:
inc r9
jmp input_loop
end_of_the_input_loop:
mov qword ptr [first_digit], 0
cmp r8, 200
jc less_than_200
sub r8, 200
mov qword ptr [first_digit], 2
less_than_200:
cmp r8, 100
jc less_than_100
sub r8, 100
mov qword ptr [first_digit], 1
less_than_100:
mov r9, 0
loop_for_dividing_r8_by_10:
cmp r8, 10
jc end_of_the_loop_for_dividing_r8_by_10
sub r8, 10
inc r9
jmp loop_for_dividing_r8_by_10
end_of_the_loop_for_dividing_r8_by_10:
shl r9, 4 # Multiply r9 by 16.
add r9, r8
mov qword ptr [binary_coded_decimal], r9
.att_syntax
)assembly");
std::cout
Он работает неправильно: < /p>
Enter a binary number containing at most 8 digits: 10111010
The number converted to decimal is: 0086
< /code>
Что происходит? Что особенно странно, так это то, что встроенная сборка успешно изменяет binary_coded_decimal Подробнее здесь: https://stackoverflow.com/questions/797 ... bled-but-d
