sinjuginas/rpm-build
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

set.c: improved rpmsetcmp main loop performance

Browse Source 
The loop is logically impeccable, but its main condition
(v1 < v1end && v2 < v2end) is somewhat redundant: in two
of the three cases, only one pointer gets advanced.  To
save instructions, the conditions are now handled within
the cases.  The loop is now a while (1) loop, a disguised
form of goto.

Also not that, when comparing Requires against Provides,
the Requires is usually sparse:

P: a b c d e f g h i j k l ...
R: a   c         h   j     ...

This means that a nested loop which skips intermediate Provides
elements towards the next Requires element may improve performance.

	while (v1 < v1end && *v1 < *v2)
	    v1++;

However, note that the first condition (v1 < v1end) is also somewhat
redundant.  This kind of boundary checking can be partially omitted if
the loop gets unrolled.  There is a better technique, however, called
the barrier: *v1end must contain the biggest element possible, so that
the trailing *v1 is never smaller than any of *v2.  The nested loop is
then becomes as simple as

	while (*v1 < *v2)
	    v1++;

callgrind annotations, 4.0.4-alt100.27:
1,899,657,916  PROGRAM TOTALS
694,132,522  decode_base62_golomb
583,376,772  rpmsetcmp
106,225,572  __GI_strcmp
102,459,314  __GI_strlen
...

callgrind annotations, this commit (rebuilt in hasher):
1,526,256,208  PROGRAM TOTALS
470,195,400  decode_base62_golomb
434,006,244  rpmsetcmp
106,137,949  __GI_strcmp
102,459,314  __GI_strlen
...

Note that rpmsetcmp also absorbs cache_decode_set and decode_delta;
the loop is now about twice as faster.
This commit is contained in:
Alexey Tourbin 2011-06-10 11:03:45 +04:00
parent 2651bb3246
commit 3ff35a310c
1 changed files with 24 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
lib/set.c
View File

@ -1092,8 +1092,9 @@ int rpmsetcmp(const char *str1, const char *str2)
if (decode_set_init(str2, &bpp2, &Mshift2) < 0)
return -4;
// make room for hash values
unsigned v1buf[decode_set_size(str1, Mshift1)], *v1 = v1buf;
unsigned v2buf[decode_set_size(str2, Mshift2)], *v2 = v2buf;
// str1 comes on behalf of provides, allocate a barrier
unsigned v1buf[decode_set_size(str1, Mshift1) + 1], *v1 = v1buf;
unsigned v2buf[decode_set_size(str2, Mshift2) + 0], *v2 = v2buf;
// decode hash values
// str1 comes on behalf of provides, decode with caching
int c1 = cache_decode_set(str1, Mshift1, v1);
@ -1116,18 +1117,32 @@ int rpmsetcmp(const char *str1, const char *str2)
int le = 1;
unsigned *v1end = v1 + c1;
unsigned *v2end = v2 + c2;
while (v1 < v1end && v2 < v2end) {
if (*v1 < *v2) {
*v1end = ~0u;
unsigned v2val = *v2;
while (1) {
if (*v1 < v2val) {
le = 0;
v1++;
while (*v1 < v2val)
v1++;
if (v1 == v1end)
break;
}
else if (*v1 > *v2) {
ge = 0;
v2++;
}
else {
if (*v1 == v2val) {
v1++;
v2++;
if (v1 == v1end)
break;
if (v2 == v2end)
break;
v2val = *v2;
}
else {
ge = 0;
v2++;
if (v2 == v2end)
break;
v2val = *v2;
}
}
// return