Re: [linux-audio-dev] Re: Fixing Freeverb and Denormals

On Sat, 22 Jan 2005 18:36:46 +1100
Erik de Castro Lopo <erikd-lad@email-addr-hidden-nerd.com> wrote:

> On Fri, 21 Jan 2005 23:08:50 +0000
> Neil Klepeis <nklepeis@email-addr-hidden> wrote:
>
> > What exactly am I looking for?
>
> I'm not too sure, but I'd know if I saw it :-).
>
> > After using g++ -S, the allpass.o file,
> > for example, starts out with:
>
> Send me the whole allpass.c file by private mail (address below) and
> I'll have a look.

OK, the code of interest is in the process() method of the allpass class.
This method is defined as an inline function in allpass.h. I therefore
wrote a small test file :

    #include "allpass.h"

    static allpass ap ;

    float test (float x)
    { return ap.process (x) ;
    }

and compiler that with:

    g++ -S -O3 test.cpp -o test.s

In the asembler output file we get this:

    .LFE6:
            .size _Z41__static_initialization_and_ <snip rest of line>
        .section .rodata.cst4,"aM",@progbits,4
        .align 4
.LC0:
        .long 201326592
        .text
        .align 2
        .p2align 4,,15

The long above (201326592) is the integer representation of the float value.

The asm code for the test function looks like this (comments added):

    .globl _Z4testf
            .type _Z4testf, @function
    _Z4testf:
    .LFB7:
            pushl %ebp
    .LCFI2:
            movl %esp, %ebp
    .LCFI3:
            subl $4, %esp
    .LCFI4:
            flds .LC0 ; load the float value to the fpu stack
            movl ap+12, %edx
            movl ap+4, %ecx
            flds 8(%ebp)
            movl (%ecx,%edx,4), %eax ; load buffer[bufidx] -> %eax
            movl %eax, -4(%ebp) ; move %eax -> bufout
            flds -4(%ebp) ; load bufout to the fpu
            fadd %st(2), %st ; new top of stack = bufout + the_const
            fstps -4(%ebp) ; store top of stack -> bufout
            flds -4(%ebp) ; load buf out to top of stack
            fsubp %st, %st(2) ; new top of stack <- bufout - the_const
                
and so on. This shows that even with -O3, the denormalisation
code is not being optimised out.

Erik

-- 
+-----------------------------------------------------------+
  Erik de Castro Lopo  nospam@email-addr-hidden-nerd.com (Yes it's valid)
+-----------------------------------------------------------+
"O'Caml ... a "language designed for smart people" if there ever
was one." -- Mike Vanier