------- Comment #19 from hubicka at gcc dot gnu dot org 2008-02-06 16:56
-------
Created an attachment (id=15108)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15108&action=view)
Complete continue heuristic patch
Hi,
this is the complete patch. With this patch we produce profile sane enough so
the internal loops are not marked cold. I will benchmark it probably tomorrow
(I want to wait for the FP changes to show separately).
It fixes the offline copy of longest_match, so we no longer have one of IV
variables at stack:
.L15:
movzbl 2(%edx), %eax
leal 2(%edx), %esi
cmpb 2(%ecx), %al
jne .L8
movzbl 3(%edx), %eax
leal 3(%edx), %esi
cmpb 3(%ecx), %al
jne .L8
movzbl 4(%edx), %eax
leal 4(%edx), %esi
cmpb 4(%ecx), %al
jne .L8
movzbl 5(%edx), %eax
leal 5(%edx), %esi
cmpb 5(%ecx), %al
jne .L8
movzbl 6(%edx), %eax
leal 6(%edx), %esi
cmpb 6(%ecx), %al
jne .L8
movzbl 7(%edx), %eax
leal 7(%edx), %esi
cmpb 7(%ecx), %al
jne .L8
leal 8(%ecx), %eax
movl %eax, %ecx
movzbl 8(%edx), %eax
cmpb (%ecx), %al
leal 8(%edx), %ebx
movl %ebx, %esi
jne .L8
cmpl %ebx, -20(%ebp)
jbe .L8
movl %ebx, %edx
movzbl 1(%edx), %eax
leal 1(%edx), %esi
cmpb 1(%ecx), %al
je .L15
Irronically this can further widen the gap in between -O2 and -O3, since the
inline copy in deflate was always allocated resonably.
Deflate codegen changes quite a lot and because function body is big I will
wait for benchmarks before trying to analyze futher.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33761
