what is the difference beetwen dw and db with dup in assembly

Question

I want to know what is the difference between this code and this code in assembly language:

a db 2 dup (1fh)
b dw 1fh,1fh

and I think that the

Answer 1

The 2nd line is equivalent to db 1fh, 0, 1fh, 0, because each arg to DW is a word-sized integer. (And x86 is little-endian)

The 1st line is equivalent to db 1fh, 1fh.

To do that with DW, use dw 1f1fh. For a very short constant, that's maybe clearer. For anything more than 2 repeats of 2 bytes, using the dup syntax is probably clearer for other humans to see that it's the same thing multiple times, without having to look carefully to check for differences.

---

If you're using an assembler which implicitly associates a size with a symbol, it matters whether you use dw or db, so if you want the symbol to be declared as a byte "variable", you need a db 1fh,1fh or a db 2 dup (1fh).

---

Think of dup as an operator where the left side is the count and the right side is the thing being repeated. Like Python's 'abc' * 3 being equivalent to "abcabcabc", or Perl's 'abc' x 4, except with the order of the operands reversed.

IDK why MASM is designed this way.

NASM syntax is times 2 db 0x1f. The times repeats a whole pseudo-instruction, or a whole instruction like times 3 imul eax, ecx.