I wrote this macro:

macro clr reg
{
xor reg, reg
}

If reg is a 64-bit register, I would like the macro to compile for the 32-bit version of that register.

For example:

clr rcx

would compile as:

xor ecx, ecx

The end result is that RCX gets zeroed as 32-bit instructions automatically zero out the upper 32-bits, but I would save a byte in memory because the REX prefix byte wouldn't need to be compiled. It might be slightly faster also.

One option would be to reverse the following macro:
http://board.flatassembler.net/topic.php?p=98098#98098

Hugh Aguilar,

There are several methods, I've thought about somewhat indirect one:

I will give that a try. This won't work for the upper registers as they have a different syntax (R13 vs R13D rather than RBP vs EBP), but I may be able to upgrade it. If not, I may be able to get by with only the lower registers.

I thought I remembered a simple way to do this, but I searched the FASM manual and couldn't find it. I may have been remembering NASM or HLA --- learning more than one assembler at a time has resulted in a mismatched jumble of information in my head!

For the most part, FASM is working out well for me --- I haven't run into much that it couldn't do.

Hugh Aguilar,

There's no reason to use it for r8d…r15d — their usage requires REX prefix too (though slightly different one).

Good point. I'll go with what you provided. Thanks.