git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Junio C Hamano <gitster@pobox.com>
To: git@vger.kernel.org
Subject: [PATCH 08/11] object: try 5-way cuckoo -- use all 20-bytes of SHA-1
Date: Thu, 11 Aug 2011 10:53:13 -0700	[thread overview]
Message-ID: <1313085196-13249-9-git-send-email-gitster@pobox.com> (raw)
In-Reply-To: <1313085196-13249-1-git-send-email-gitster@pobox.com>

Signed-off-by: Junio C Hamano <gitster@pobox.com>
---
../+v/ced878c983ddc9a256b22535aa8af006529ec3d0/git-pack-objects
Counting objects: 2139209, done.
31.66user 2.14system 0:33.91elapsed 99%CPU (0avgtext+0avgdata 2874176maxresident)k
0inputs+0outputs (0major+225342minor)pagefaults 0swaps
Counting objects: 2139209, done.
31.79user 2.03system 0:33.93elapsed 99%CPU (0avgtext+0avgdata 2875808maxresident)k
0inputs+0outputs (0major+225444minor)pagefaults 0swaps
Counting objects: 2139209, done.
31.80user 1.94system 0:33.84elapsed 99%CPU (0avgtext+0avgdata 2875792maxresident)k
0inputs+0outputs (0major+225443minor)pagefaults 0swaps
---
 object.c |   43 ++++++++++++++++++++++++++++++-------------
 1 files changed, 30 insertions(+), 13 deletions(-)

diff --git a/object.c b/object.c
index 7624c48..c777520 100644
--- a/object.c
+++ b/object.c
@@ -43,27 +43,27 @@ struct object *get_indexed_object(unsigned int idx)
 	return obj_hash[idx];
 }
 
-static unsigned int hash_val(const unsigned char *sha1)
-{
-	unsigned int hash;
-	memcpy(&hash, sha1, sizeof(unsigned int));
-	return hash;
-}
-
 
-#define H1(sha1) (hash_val(sha1) % obj_hash_size)
-#define H2(sha1) (hash_val((sha1) + sizeof(unsigned int)) % obj_hash_size)
+#define H(hv,ix) ((hv[ix]) & (obj_hash_size-1))
 
 struct object *lookup_object(const unsigned char *sha1)
 {
 	struct object *obj;
+	unsigned int hashval[5];
 
 	if (!obj_hash)
 		return NULL;
 
-	if ((obj = obj_hash[H1(sha1)]) && !hashcmp(sha1, obj->sha1))
+	memcpy(hashval, sha1, 20);
+	if ((obj = obj_hash[H(hashval, 0)]) && !hashcmp(sha1, obj->sha1))
+		return obj;
+	if ((obj = obj_hash[H(hashval, 1)]) && !hashcmp(sha1, obj->sha1))
+		return obj;
+	if ((obj = obj_hash[H(hashval, 2)]) && !hashcmp(sha1, obj->sha1))
 		return obj;
-	if ((obj = obj_hash[H2(sha1)]) && !hashcmp(sha1, obj->sha1))
+	if ((obj = obj_hash[H(hashval, 3)]) && !hashcmp(sha1, obj->sha1))
+		return obj;
+	if ((obj = obj_hash[H(hashval, 4)]) && !hashcmp(sha1, obj->sha1))
 		return obj;
 	return NULL;
 }
@@ -84,13 +84,30 @@ static struct object *insert_obj_hash(struct object *obj)
 	for (loop = obj_hash_size / 2; 0 <= loop; loop--) {
 		struct object *tmp_obj;
 		unsigned int ix;
+		unsigned int hashval[5];
 
-		ix = H1(obj->sha1);
+		memcpy(hashval, obj->sha1, 20);
+		ix = H(hashval, 0);
+		if (!obj_hash[ix]) {
+			obj_hash[ix] = obj;
+			return NULL;
+		}
+		ix = H(hashval, 1);
+		if (!obj_hash[ix]) {
+			obj_hash[ix] = obj;
+			return NULL;
+		}
+		ix = H(hashval, 2);
+		if (!obj_hash[ix]) {
+			obj_hash[ix] = obj;
+			return NULL;
+		}
+		ix = H(hashval, 3);
 		if (!obj_hash[ix]) {
 			obj_hash[ix] = obj;
 			return NULL;
 		}
-		ix = H2(obj->sha1);
+		ix = H(hashval, 4);
 		tmp_obj = obj_hash[ix];
 		obj_hash[ix] = obj;
 		if (!tmp_obj)
-- 
1.7.6.433.g1421f

  parent reply	other threads:[~2011-08-11 17:54 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-08-11 17:53 [PATCH 00/11] Micro-optimizing lookup_object() Junio C Hamano
2011-08-11 17:53 ` [PATCH 01/11] object.c: code movement for readability Junio C Hamano
2011-08-11 17:53 ` [PATCH 02/11] object.c: remove duplicated code for object hashing Junio C Hamano
2011-08-11 17:53 ` [PATCH 03/11] pack-objects --count-only Junio C Hamano
2011-08-11 17:53 ` [PATCH 04/11] object: next_size() helper for readability Junio C Hamano
2011-08-11 17:53 ` [PATCH 05/11] object hash: we know the table size is a power of two Junio C Hamano
2011-08-11 17:53 ` [PATCH 06/11] object: growing the hash-table more aggressively does not help much Junio C Hamano
2011-08-11 17:53 ` [PATCH 07/11] object: try naive cuckoo hashing Junio C Hamano
2011-08-11 17:53 ` Junio C Hamano [this message]
2011-08-11 17:53 ` [PATCH 09/11] object: try 4-way cuckoo Junio C Hamano
2011-08-11 17:53 ` [PATCH 10/11] object: try 3-way cuckoo Junio C Hamano
2011-08-11 17:53 ` [PATCH 11/11] object: try 2-way cuckoo again Junio C Hamano
2011-08-11 23:33 ` [FFT/PATCH 12/11] object.c: make object hash implementation more opaque Junio C Hamano
2011-08-12 15:59 ` git_checkattr() is inefficient when repeated [Re: [PATCH 00/11] Micro-optimizing lookup_object()] Thomas Rast
2011-08-15 23:19   ` Junio C Hamano

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1313085196-13249-9-git-send-email-gitster@pobox.com \
    --to=gitster@pobox.com \
    --cc=git@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).