IntelVsATnTSyntax

There are few major difference between Intel and AT&T syntax of assembly language:

1) In AT&T mnemonic, register names must be in lowercase. Whether in Intel syntax both lowercase and uppercase register names are acceptable.

2) Register names in AT&T syntax must preceed by a '%' symbol. AX must be mentioned as %ax, RSP asn %rsp etc.

3) Every opcode that take oparands, must end with a suffix that indicate the size of the operands in AT&T syntax. For example: "push rsp" in
   Intel syntax becomes "pushq %rsp" in AT&T syntax. Note: opcodes that don't take opcodes, like call, ret etc don't take such suffix.

4) The order of source and destination operand is reverse in AT&T syntax with respect to Intel syntax. For Intel syntax this relation have
   been:
   		<Opcode> <dest>, <source>

   In AT&T syntax it becomes:
	
		<Opcode> <source>, <dest>

   So, "mov rbx, rsp" in Intel syntax becomes "movq %rsp, %rbx" in AT&T syntax.

5) In AT&T syntax, immediate operands always preceed by '$'. So, "mov rax, 256" in Intel syntax becomes "movq $356, %rax". When referenced
   message string names are prefixed by '$' as well.

6) Memory adressing is completely different in these two syntaxes. In Intel Syntax the memory addressing is the straight forward mathematical
formula:
	<base> + (<Index>*<Scale>)+<displacement>

AT&T handles this with a special syntax:
	+/-<displacement>(<base>, <index>, <scale>)

Thus "rbp-16" in Intel syntax is "-16(%rbp)" in AT&T syntax. Similarly, "mov rcx, [rbx+rdi*2+28]" becomes "movq 28(%rbx, %rdi, $2), %rcx".
Similarly "mov qword rax, [rbx-4]" becomes "movq -4(%rbx), %rax".

7) In AT&T syntax comment delimiter is '#' instead of ';' in Intel syntax.

To end, assembly program written in AT&T syntax has the extension ".s", where those written in intel syntax has the extension ".asm".