How to convert an integer to a floating point value in x86 ASM?

问题

I need to multiply an integer (two's compliment) by a floating point constant. Here is what I have:

.data
pi  dd 3.14
int dd 0ah


.code
fld pi

???

fmul ST(1), ST

How can I convert int to a floating point value for multiplying against pi?

回答1:

You need the fild instruction. Here's one reference: http://www.website.masmforum.com/tutorials/fptute/fpuchap5.htm

回答2:

Using the x86 FPU is really outdated now and way slower than SSE/AVX/etc.

Better to use at least SSE with cvtdq2psdoc

ex:

.data
int dd 0ah
pi  dd 3.14

.code
movss xmm0, [data] ; load 32bit data into xmm0
cvtdq2ps xmm0, xmm0 ; convert 4 float values to integers, we use only lowest float/integer
mulss xmm0, [pi]

来源：https://stackoverflow.com/questions/3027424/how-to-convert-an-integer-to-a-floating-point-value-in-x86-asm