xdiff/xprepare: use a smaller sample size for histogram diff

For histogram diff, we can afford a smaller sample size and thus a
poorer estimate of the number of lines, as the hash table (rhash) won't
be filled up/grown. This is safe as the final count of lines (xdf.nrecs)
will be updated correctly anyway by xdl_prepare_ctx().

This gives us a small boost in performance.

Signed-off-by: Tay Ray Chuan <rctay89@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This commit is contained in:
Tay Ray Chuan 2011-07-12 14:10:27 +08:00 committed by Junio C Hamano
parent 9f37c27593
commit 86abba8015
3 changed files with 17 additions and 10 deletions

View File

@ -26,6 +26,8 @@
#define XDL_KPDIS_RUN 4 #define XDL_KPDIS_RUN 4
#define XDL_MAX_EQLIMIT 1024 #define XDL_MAX_EQLIMIT 1024
#define XDL_SIMSCAN_WINDOW 100 #define XDL_SIMSCAN_WINDOW 100
#define XDL_GUESS_NLINES1 256
#define XDL_GUESS_NLINES2 20
typedef struct s_xdlclass { typedef struct s_xdlclass {
@ -239,11 +241,20 @@ static void xdl_free_ctx(xdfile_t *xdf) {
int xdl_prepare_env(mmfile_t *mf1, mmfile_t *mf2, xpparam_t const *xpp, int xdl_prepare_env(mmfile_t *mf1, mmfile_t *mf2, xpparam_t const *xpp,
xdfenv_t *xe) { xdfenv_t *xe) {
long enl1, enl2; long enl1, enl2, sample;
xdlclassifier_t cf; xdlclassifier_t cf;
enl1 = xdl_guess_lines(mf1) + 1; /*
enl2 = xdl_guess_lines(mf2) + 1; * For histogram diff, we can afford a smaller sample size and
* thus a poorer estimate of the number of lines, as the hash
* table (rhash) won't be filled up/grown. The number of lines
* (nrecs) will be updated correctly anyway by
* xdl_prepare_ctx().
*/
sample = xpp->flags & XDF_HISTOGRAM_DIFF ? XDL_GUESS_NLINES2 : XDL_GUESS_NLINES1;
enl1 = xdl_guess_lines(mf1, sample) + 1;
enl2 = xdl_guess_lines(mf2, sample) + 1;
if (!(xpp->flags & XDF_HISTOGRAM_DIFF) && if (!(xpp->flags & XDF_HISTOGRAM_DIFF) &&
xdl_init_classifier(&cf, enl1 + enl2 + 1, xpp->flags) < 0) { xdl_init_classifier(&cf, enl1 + enl2 + 1, xpp->flags) < 0) {

View File

@ -24,10 +24,6 @@
#define XDL_GUESS_NLINES 256
long xdl_bogosqrt(long n) { long xdl_bogosqrt(long n) {
long i; long i;
@ -159,12 +155,12 @@ void *xdl_cha_next(chastore_t *cha) {
} }
long xdl_guess_lines(mmfile_t *mf) { long xdl_guess_lines(mmfile_t *mf, long sample) {
long nl = 0, size, tsize = 0; long nl = 0, size, tsize = 0;
char const *data, *cur, *top; char const *data, *cur, *top;
if ((cur = data = xdl_mmfile_first(mf, &size)) != NULL) { if ((cur = data = xdl_mmfile_first(mf, &size)) != NULL) {
for (top = data + size; nl < XDL_GUESS_NLINES;) { for (top = data + size; nl < sample;) {
if (cur >= top) { if (cur >= top) {
tsize += (long) (cur - data); tsize += (long) (cur - data);
if (!(cur = data = xdl_mmfile_next(mf, &size))) if (!(cur = data = xdl_mmfile_next(mf, &size)))

View File

@ -33,7 +33,7 @@ void xdl_cha_free(chastore_t *cha);
void *xdl_cha_alloc(chastore_t *cha); void *xdl_cha_alloc(chastore_t *cha);
void *xdl_cha_first(chastore_t *cha); void *xdl_cha_first(chastore_t *cha);
void *xdl_cha_next(chastore_t *cha); void *xdl_cha_next(chastore_t *cha);
long xdl_guess_lines(mmfile_t *mf); long xdl_guess_lines(mmfile_t *mf, long sample);
int xdl_recmatch(const char *l1, long s1, const char *l2, long s2, long flags); int xdl_recmatch(const char *l1, long s1, const char *l2, long s2, long flags);
unsigned long xdl_hash_record(char const **data, char const *top, long flags); unsigned long xdl_hash_record(char const **data, char const *top, long flags);
unsigned int xdl_hashbits(unsigned int size); unsigned int xdl_hashbits(unsigned int size);