Hi --
I'm seeing register allocation problems and code size increases
with gcc-4.6.2 (and gcc-head) compared with older (gcc-4.1.2).
Both are compiled using -O3.
One test case that I have has a long series of nested if's
each with the same comparison and similar computation.
if (n<max_no){
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p);
if (n<max_no){
. . . ~20 levels of nesting
<more computations with 'cp' and 'p'>
. . . }}}
Gcc-4.6.2 generates many blocks like the following:
lwi r28,r1,68 -- load into dead reg
lwi r31,r1,140 -- load p from stack
lbui r28,r31,0
rsubk r31,r28,r19
lbui r31,r31,0
addk r29,r29,r31
swi r31,r1,308
lwi r31,r1,428 -- load of max_no from stack
cmp r28,r31,r29 -- n in r29
bgeid r28,$L46
gcc-4.1.2 generates the following:
lbui r3,r26,3
rsubk r3,r3,r19
lbui r3,r3,0
addk r30,r30,r3
swi r3,r1,80
cmp r18,r9,r30 -- max_no in r9, n in r30
bgei r18,$L6
gcc-4.6.2 (and gcc-head) load max_no from the stack in each block.
There also are extra loads into r28 (which is not used) and r31 at
the start of each block. Only r28, r29, and r31 are used.
I'm having a hard time telling what is happening or why. The
IRA dump has this line:
Ignoring reg 772, has equiv memory
where pseudo 772 is loaded with max_no early in the function.
The reload dump has
Reloads for insn # 254
Reload 0: reload_in (SI) = (reg/v:SI 722 [ max_no ])
GR_REGS, RELOAD_FOR_INPUT (opnum = 1)
reload_in_reg: (reg/v:SI 722 [ max_no ])
reload_reg_rtx: (reg:SI 31 r31)
and similar for each of the other insns using 722.
This is followed by
Spilling for insn 254.
Using reg 31 for reload 0
for each insn using pseudo 722.
Any idea what is going on?
--
Michael Eager ea...@eagercon.com
1960 Park Blvd., Palo Alto, CA 94306 650-325-8077
#if 0
mb-gcc -O3 -mhard-float -fdump-rtl-all -c s.c -save-temps
#endif
typedef unsigned char uchar;
typedef struct {int x,y,info, dx, dy, I;} CORNER_LIST[15000];
susan_corners(in,r,bp,max_no,corner_list,x_size,y_size)
uchar *in, *bp;
int *r, max_no, x_size, y_size;
CORNER_LIST corner_list;
{
int n,x,y,sq,xx,yy,
i,j,*cgx,*cgy;
float divide;
uchar c,*p,*cp;
for (i=5;i<y_size-5;i++)
for (j=5;j<x_size-5;j++)
{
n=100;
p=in + (i-3)*x_size + j - 1;
cp=bp + in[i*x_size+j];
n+=*(cp-*p++);
n+=*(cp-*p++);
n+=*(cp-*p);
p+=x_size-3;
n+=*(cp-*p++);
n+=*(cp-*p++);
n+=*(cp-*p++);
n+=*(cp-*p++);
n+=*(cp-*p);
p+=x_size-5;
n+=*(cp-*p++);
n+=*(cp-*p++);
n+=*(cp-*p++);
n+=*(cp-*p++);
n+=*(cp-*p++);
n+=*(cp-*p++);
n+=*(cp-*p);
p+=x_size-6;
n+=*(cp-*p++);
n+=*(cp-*p++);
n+=*(cp-*p);
if (n<max_no){
p+=2;
n+=*(cp-*p++);
#if 1
if (n<max_no){
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p);
if (n<max_no){
p+=x_size-6;
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p);
if (n<max_no){
p+=x_size-5;
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p);
if (n<max_no){
p+=x_size-3;
n+=*(cp-*p++);
#endif
if (n<max_no){
n+=*(cp-*p++);
if (n<max_no){
n+=*(cp-*p);
if (n<max_no)
{
x=0;y=0;
p=in + (i-3)*x_size + j - 1;
c=*(cp-*p++);x-=c;y-=3*c;
c=*(cp-*p++);y-=3*c;
c=*(cp-*p);x+=c;y-=3*c;
p+=x_size-3;
c=*(cp-*p++);x-=2*c;y-=2*c;
c=*(cp-*p++);x-=c;y-=2*c;
c=*(cp-*p++);y-=2*c;
c=*(cp-*p++);x+=c;y-=2*c;
c=*(cp-*p);x+=2*c;y-=2*c;
p+=x_size-5;
c=*(cp-*p++);x-=3*c;y-=c;
c=*(cp-*p++);x-=2*c;y-=c;
c=*(cp-*p++);x-=c;y-=c;
c=*(cp-*p++);y-=c;
c=*(cp-*p++);x+=c;y-=c;
c=*(cp-*p++);x+=2*c;y-=c;
c=*(cp-*p);x+=3*c;y-=c;
p+=x_size-6;
c=*(cp-*p++);x-=3*c;
c=*(cp-*p++);x-=2*c;
c=*(cp-*p);x-=c;
p+=2;
c=*(cp-*p++);x+=c;
c=*(cp-*p++);x+=2*c;
c=*(cp-*p);x+=3*c;
p+=x_size-6;
c=*(cp-*p++);x-=3*c;y+=c;
c=*(cp-*p++);x-=2*c;y+=c;
c=*(cp-*p++);x-=c;y+=c;
c=*(cp-*p++);y+=c;
c=*(cp-*p++);x+=c;y+=c;
c=*(cp-*p++);x+=2*c;y+=c;
c=*(cp-*p);x+=3*c;y+=c;
p+=x_size-5;
c=*(cp-*p++);x-=2*c;y+=2*c;
c=*(cp-*p++);x-=c;y+=2*c;
c=*(cp-*p++);y+=2*c;
c=*(cp-*p++);x+=c;y+=2*c;
c=*(cp-*p);x+=2*c;y+=2*c;
p+=x_size-3;
c=*(cp-*p++);x-=c;y+=3*c;
c=*(cp-*p++);y+=3*c;
c=*(cp-*p);x+=c;y+=3*c;
xx=x*x;
yy=y*y;
sq=xx+yy;
if ( sq > ((n*n)/2) )
{
#if 1
if(yy<xx)
{
divide=(float)y/(float)abs(x);
sq=abs(x)/x;
sq=*(cp-in[(i+( (divide) < 0 ? ((int)(divide-0.5)) : ((int)(divide+0.5)) ))*x_size+j+sq]) +
*(cp-in[(i+( (2*divide) < 0 ? ((int)(2*divide-0.5)) : ((int)(2*divide+0.5)) ))*x_size+j+2*sq]) +
*(cp-in[(i+( (3*divide) < 0 ? ((int)(3*divide-0.5)) : ((int)(3*divide+0.5)) ))*x_size+j+3*sq]);
}
else
{
divide=(float)x/(float)abs(y);
sq=abs(y)/y;
sq=*(cp-in[(i+sq)*x_size+j+( (divide) < 0 ? ((int)(divide-0.5)) : ((int)(divide+0.5)) )]) +
*(cp-in[(i+2*sq)*x_size+j+( (2*divide) < 0 ? ((int)(2*divide-0.5)) : ((int)(2*divide+0.5)) )]) +
*(cp-in[(i+3*sq)*x_size+j+( (3*divide) < 0 ? ((int)(3*divide-0.5)) : ((int)(3*divide+0.5)) )]);
}
if(sq>290)
{
r[i*x_size+j] = max_no-n;
cgx[i*x_size+j] = (51*x)/n;
cgy[i*x_size+j] = (51*y)/n;
}
#endif
}
}
}
}
#if 1
}
}
}
}
}
}
}
}
}
}
}
}
}
}
}
#endif
}
}
#if 0
n=0;
for (i=5;i<y_size-5;i++)
for (j=5;j<x_size-5;j++) {
x = r[i*x_size+j];
if (x>0) {
if (
(x>r[(i-3)*x_size+j-3]) &&
(x>r[(i-3)*x_size+j-2]) &&
(x>r[(i-3)*x_size+j-1]) &&
(x>r[(i-3)*x_size+j ]) &&
(x>r[(i-3)*x_size+j+1]) &&
(x>r[(i-3)*x_size+j+2]) &&
(x>r[(i-3)*x_size+j+3]) &&
(x>r[(i-2)*x_size+j-3]) &&
(x>r[(i-2)*x_size+j-2]) &&
(x>r[(i-2)*x_size+j-1]) &&
(x>r[(i-2)*x_size+j ]) &&
(x>r[(i-2)*x_size+j+1]) &&
(x>r[(i-2)*x_size+j+2]) &&
(x>r[(i-2)*x_size+j+3]) &&
(x>r[(i-1)*x_size+j-3]) &&
(x>r[(i-1)*x_size+j-2]) &&
(x>r[(i-1)*x_size+j-1]) &&
(x>r[(i-1)*x_size+j ]) &&
(x>r[(i-1)*x_size+j+1]) &&
(x>r[(i-1)*x_size+j+2]) &&
(x>r[(i-1)*x_size+j+3]) &&
(x>r[(i)*x_size+j-3]) &&
(x>r[(i)*x_size+j-2]) &&
(x>r[(i)*x_size+j-1]) &&
(x>=r[(i)*x_size+j+1]) &&
(x>=r[(i)*x_size+j+2]) &&
(x>=r[(i)*x_size+j+3]) &&
(x>=r[(i+1)*x_size+j-3]) &&
(x>=r[(i+1)*x_size+j-2]) &&
(x>=r[(i+1)*x_size+j-1]) &&
(x>=r[(i+1)*x_size+j ]) &&
(x>=r[(i+1)*x_size+j+1]) &&
(x>=r[(i+1)*x_size+j+2]) &&
(x>=r[(i+1)*x_size+j+3]) &&
(x>=r[(i+2)*x_size+j-3]) &&
(x>=r[(i+2)*x_size+j-2]) &&
(x>=r[(i+2)*x_size+j-1]) &&
(x>=r[(i+2)*x_size+j ]) &&
(x>=r[(i+2)*x_size+j+1]) &&
(x>=r[(i+2)*x_size+j+2]) &&
(x>=r[(i+2)*x_size+j+3]) &&
(x>=r[(i+3)*x_size+j-3]) &&
(x>=r[(i+3)*x_size+j-2]) &&
(x>=r[(i+3)*x_size+j-1]) &&
(x>=r[(i+3)*x_size+j ]) &&
(x>=r[(i+3)*x_size+j+1]) &&
(x>=r[(i+3)*x_size+j+2]) &&
(x>=r[(i+3)*x_size+j+3])
)
{
corner_list[n].info=0;
corner_list[n].x=j;
corner_list[n].y=i;
corner_list[n].dx=cgx[i*x_size+j];
corner_list[n].dy=cgy[i*x_size+j];
corner_list[n].I=in[i*x_size+j];
n++;
if(n==15000){
}
}
}
}
#endif
corner_list[n].info=7;
}
