Changeset: a9338d708269 for MonetDB
URL: http://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=a9338d708269
Modified Files:
gdk/gdk_bbp.c
gdk/gdk_join.c
gdk/gdk_private.h
gdk/gdk_search.c
gdk/gdk_select.c
gdk/gdk_unique.c
monetdb5/extras/rdf/rdftypes.c
Branch: rdf
Log Message:
Disable parent hash, persistent hash
diffs (truncated from 330 to 300 lines):
diff --git a/gdk/gdk_bbp.c b/gdk/gdk_bbp.c
--- a/gdk/gdk_bbp.c
+++ b/gdk/gdk_bbp.c
@@ -3839,11 +3839,23 @@ BBPdiskscan(const char *parent)
BAT *b = getdesc(bid);
delete = (b == NULL || !b->T->vheap ||
b->batCopiedtodisk == 0);
} else if (strncmp(p + 1, "hhash", 5) == 0) {
+#ifdef PERSISTENTHASH
BAT *b = getdesc(bid);
delete = b == NULL;
+ if (!delete)
+ b->H->hash = (Hash *) 1;
+#else
+ delete = TRUE;
+#endif
} else if (strncmp(p + 1, "thash", 5) == 0) {
+#ifdef PERSISTENTHASH
BAT *b = getdesc(bid);
delete = b == NULL;
+ if (!delete)
+ b->T->hash = (Hash *) 1;
+#else
+ delete = TRUE;
+#endif
} else if (strncmp(p + 1, "himprints", 9) == 0) {
BAT *b = getdesc(bid);
delete = b == NULL;
diff --git a/gdk/gdk_join.c b/gdk/gdk_join.c
--- a/gdk/gdk_join.c
+++ b/gdk/gdk_join.c
@@ -1843,19 +1843,36 @@ hashjoin(BAT *r1, BAT *r2, BAT *l, BAT *
return GDK_SUCCEED;
}
+ rl = BUNfirst(r);
+#ifndef DISABLE_PARENT_HASH
if (VIEWtparent(r)) {
BAT *b = BBPdescriptor(-VIEWtparent(r));
- rl = (BUN) ((r->T->heap.base - b->T->heap.base) >> r->T->shift)
+ BUNfirst(r);
- r = b;
- } else {
- rl = BUNfirst(r);
+ if (b->batPersistence == PERSISTENT || BATcheckhash(b)) {
+ /* only use parent's hash if it is persistent
+ * or already has a hash */
+ ALGODEBUG
+ fprintf(stderr, "#hashjoin(%s#"BUNFMT"): "
+ "using parent(%s#"BUNFMT") for hash\n",
+ BATgetId(r), BATcount(r),
+ BATgetId(b), BATcount(b));
+ rl = (BUN) ((r->T->heap.base - b->T->heap.base) >>
r->T->shift) + BUNfirst(r);
+ r = b;
+ } else {
+ ALGODEBUG
+ fprintf(stderr, "#hashjoin(%s#"BUNFMT"): not "
+ "using parent(%s#"BUNFMT") for hash\n",
+ BATgetId(r), BATcount(r),
+ BATgetId(b), BATcount(b));
+ }
}
+#endif
rh = rl + rend;
rl += rstart;
rseq += rstart;
if (BAThash(r, 0) == GDK_FAIL)
goto bailout;
+
ri = bat_iterator(r);
nrcand = (BUN) (rcandend - rcand);
@@ -2871,7 +2888,9 @@ BATsubjoin(BAT **r1p, BAT **r2p, BAT *l,
BUN lcount, rcount, lpcount, rpcount;
BUN lsize, rsize;
int lhash, rhash;
+#ifndef DISABLE_PARENT_HASH
bat lparent, rparent;
+#endif
int swap;
size_t mem_size;
@@ -2912,19 +2931,26 @@ BATsubjoin(BAT **r1p, BAT **r2p, BAT *l,
rsize = (BUN) (BATcount(r) * (Tsize(r) + (r->T->vheap ?
r->T->vheap->size : 0) + 2 * sizeof(BUN)));
mem_size = GDK_mem_maxsize / (GDKnr_threads ? GDKnr_threads : 1);
+#ifndef DISABLE_PARENT_HASH
lparent = VIEWtparent(l);
- rparent = VIEWtparent(r);
if (lparent) {
lpcount = BATcount(BBPdescriptor(lparent));
lhash = BATcheckhash(l) ||
BATcheckhash(BBPdescriptor(-lparent));
- } else {
+ } else
+#endif
+ {
lpcount = BATcount(l);
lhash = BATcheckhash(l);
}
+
+#ifndef DISABLE_PARENT_HASH
+ rparent = VIEWtparent(r);
if (rparent) {
rpcount = BATcount(BBPdescriptor(rparent));
rhash = BATcheckhash(r) ||
BATcheckhash(BBPdescriptor(-rparent));
- } else {
+ } else
+#endif
+ {
rpcount = BATcount(r);
rhash = BATcheckhash(r);
}
@@ -2963,21 +2989,33 @@ BATsubjoin(BAT **r1p, BAT **r2p, BAT *l,
* large (i.e. prefer hash over binary search, but
* only if the hash table doesn't cause thrashing) */
return mergejoin(r1, r2, l, r, sl, sr, nil_matches, 0, 0, 0);
- } else if ((l->batPersistence == PERSISTENT ||
- (lparent != 0 &&
- BBPquickdesc(abs(lparent), 0)->batPersistence ==
PERSISTENT)) &&
- !(r->batPersistence == PERSISTENT ||
- (rparent != 0 &&
- BBPquickdesc(abs(rparent), 0)->batPersistence ==
PERSISTENT))) {
+ } else if ((l->batPersistence == PERSISTENT
+#ifndef DISABLE_PARENT_HASH
+ || (lparent != 0 &&
+ BBPquickdesc(abs(lparent), 0)->batPersistence ==
PERSISTENT)
+#endif
+ ) &&
+ !(r->batPersistence == PERSISTENT
+#ifndef DISABLE_PARENT_HASH
+ || (rparent != 0 &&
+ BBPquickdesc(abs(rparent), 0)->batPersistence ==
PERSISTENT)
+#endif
+ )) {
/* l (or its parent) is persistent and r is not,
* create hash on l since it may be reused */
swap = 1;
- } else if (!(l->batPersistence == PERSISTENT ||
- (lparent != 0 &&
- BBPquickdesc(abs(lparent), 0)->batPersistence ==
PERSISTENT)) &&
- (r->batPersistence == PERSISTENT ||
- (rparent != 0 &&
- BBPquickdesc(abs(rparent), 0)->batPersistence ==
PERSISTENT))) {
+ } else if (!(l->batPersistence == PERSISTENT
+#ifndef DISABLE_PARENT_HASH
+ || (lparent != 0 &&
+ BBPquickdesc(abs(lparent), 0)->batPersistence ==
PERSISTENT)
+#endif
+ ) &&
+ (r->batPersistence == PERSISTENT
+#ifndef DISABLE_PARENT_HASH
+ || (rparent != 0 &&
+ BBPquickdesc(abs(rparent), 0)->batPersistence ==
PERSISTENT)
+#endif
+ )) {
/* l (and its parent) is not persistent but r (or its
* parent) is, create hash on r since it may be
* reused */
@@ -2986,6 +3024,7 @@ BATsubjoin(BAT **r1p, BAT **r2p, BAT *l,
/* no hashes, not sorted, create hash on smallest BAT */
swap = 1;
}
+
if (swap) {
return hashjoin(r2, r1, r, l, sr, sl, nil_matches, 0, 0, 0);
} else {
diff --git a/gdk/gdk_private.h b/gdk/gdk_private.h
--- a/gdk/gdk_private.h
+++ b/gdk/gdk_private.h
@@ -12,6 +12,8 @@
#error this file should not be included outside its source directory
#endif
+#define DISABLE_PARENT_HASH 1
+
#include "gdk_system_private.h"
enum heaptype {
diff --git a/gdk/gdk_search.c b/gdk/gdk_search.c
--- a/gdk/gdk_search.c
+++ b/gdk/gdk_search.c
@@ -243,7 +243,11 @@ BATcheckhash(BAT *b)
if ((h = GDKmalloc(sizeof(*h))) != NULL &&
read(fd, hdata, sizeof(hdata)) ==
sizeof(hdata) &&
- hdata[0] == (((size_t) 1 << 24) |
HASH_VERSION) &&
+ hdata[0] == (
+#ifdef PERSISTENTHASH
+ ((size_t) 1 << 24) |
+#endif
+ HASH_VERSION) &&
hdata[4] == (size_t) BATcount(b) &&
fstat(fd, &st) == 0 &&
st.st_size >= (off_t) (hp->size = hp->free
= (hdata[1] + hdata[2]) * hdata[3] + HASH_HEADER_SIZE * SIZEOF_SIZE_T) &&
@@ -320,6 +324,8 @@ BAThash(BAT *b, BUN masksize)
}
return GDK_SUCCEED;
}
+
+
MT_lock_set(&GDKhashLock(abs(b->batCacheid)), "BAThash");
if (b->T->hash == NULL) {
unsigned int tpe = ATOMbasetype(b->ttype);
@@ -331,7 +337,9 @@ BAThash(BAT *b, BUN masksize)
const char *nme = BBP_physical(b->batCacheid);
const char *ext = b->batCacheid > 0 ? "thash" : "hhash";
BATiter bi = bat_iterator(b);
+#ifdef PERSISTENTHASH
int fd;
+#endif
ALGODEBUG fprintf(stderr, "#BAThash: create hash(" BUNFMT
");\n", BATcount(b));
if ((hp = GDKzalloc(sizeof(*hp))) == NULL ||
@@ -503,6 +511,7 @@ BAThash(BAT *b, BUN masksize)
}
break;
}
+#ifdef PERSISTENTHASH
if ((BBP_status(b->batCacheid) & BBPEXISTING) &&
HEAPsave(hp, nme, ext) == 0 &&
(fd = GDKfdlocate(hp->farmid, nme, "rb+", ext)) >= 0) {
@@ -522,6 +531,7 @@ BAThash(BAT *b, BUN masksize)
close(fd);
} else
ALGODEBUG fprintf(stderr, "#BAThash: NOT persisting
hash %d\n", b->batCacheid);
+#endif
b->T->hash = h;
t1 = GDKusec();
ALGODEBUG fprintf(stderr, "#BAThash: hash construction " LLFMT
" usec\n", t1 - t0);
diff --git a/gdk/gdk_select.c b/gdk/gdk_select.c
--- a/gdk/gdk_select.c
+++ b/gdk/gdk_select.c
@@ -1599,18 +1599,27 @@ BATsubselect(BAT *b, BAT *s, const void
* persistent and the total size wouldn't be too large; check
* for existence of hash last since that may involve I/O */
hash = equi &&
- (((b->batPersistence == PERSISTENT ||
- (parent != 0 &&
- BBPquickdesc(abs(parent),0)->batPersistence == PERSISTENT))
&&
+ (((b->batPersistence == PERSISTENT
+#ifndef DISABLE_PARENT_HASH
+ || (parent != 0 &&
+ BBPquickdesc(abs(parent),0)->batPersistence == PERSISTENT)
+#endif
+ ) &&
(size_t) ATOMsize(b->ttype) >= sizeof(BUN) / 4 &&
BATcount(b) * (ATOMsize(b->ttype) + 2 * sizeof(BUN)) <
GDK_mem_maxsize / 2) ||
- (BATcheckhash(b) ||
- (parent != 0 &&
- BATcheckhash(BBPdescriptor(-parent)))));
+ (BATcheckhash(b)
+#ifndef DISABLE_PARENT_HASH
+ || (parent != 0 &&
+ BATcheckhash(BBPdescriptor(-parent)))
+#endif
+ ));
if (hash &&
estimate == BUN_NONE &&
- !BATcheckhash(b) &&
- (parent == 0 || !BATcheckhash(BBPdescriptor(-parent)))) {
+ !BATcheckhash(b)
+#ifndef DISABLE_PARENT_HASH
+ && (parent == 0 || !BATcheckhash(BBPdescriptor(-parent)))
+#endif
+ ) {
/* no exact result size, but we need estimate to choose
* between hash- & scan-select
* (if we already have a hash, it's a no-brainer: we
diff --git a/gdk/gdk_unique.c b/gdk/gdk_unique.c
--- a/gdk/gdk_unique.c
+++ b/gdk/gdk_unique.c
@@ -41,7 +41,9 @@ BATsubunique(BAT *b, BAT *s)
BUN hb;
BATiter bi;
int (*cmp)(const void *, const void *);
+#ifndef DISABLE_PARENT_HASH
bat parent;
+#endif
BATcheck(b, "BATsubunique", NULL);
if (b->tkey || BATcount(b) <= 1 || BATtdense(b)) {
@@ -243,9 +245,12 @@ BATsubunique(BAT *b, BAT *s)
seen = NULL;
} else if (BATcheckhash(b) ||
(b->batPersistence == PERSISTENT &&
- BAThash(b, 0) == GDK_SUCCEED) ||
- ((parent = VIEWtparent(b)) != 0 &&
- BATcheckhash(BBPdescriptor(-parent)))) {
+ BAThash(b, 0) == GDK_SUCCEED)
+#ifndef DISABLE_PARENT_HASH
+ || ((parent = VIEWtparent(b)) != 0 &&
+ BATcheckhash(BBPdescriptor(-parent)))
+#endif
+ ) {
BUN lo;
oid seq;
@@ -257,12 +262,15 @@ BATsubunique(BAT *b, BAT *s)
s ? BATgetId(s) : "NULL",
s ? BATcount(s) : 0);
seq = b->hseqbase;
_______________________________________________
checkin-list mailing list
checkin-list@monetdb.org
https://www.monetdb.org/mailman/listinfo/checkin-list
