[PATCH] improve diff-delta with sparse and/or repetitive data

Nicolas Pitre· May 3, 2006, 03:31 UTC · lore

It is useless to preserve multiple hash entries for consecutive blocks with the same hash. Keeping only the first one will allow for matching the longest string of identical bytes while subsequent blocks will only allow for shorter matches. The backward matching code will match the end of it as necessary.

This improves both performances (no repeated string compare with long successions of identical bytes, or even small group of bytes), as well as compression (less likely to need random hash bucket entry culling), especially with sparse files.

With well behaved data sets this patch doesn't change much.

Signed-off-by: Nicolas Pitre <nico@cam.org>

---

Show changes to diff-delta.c +27 −13

diff --git a/diff-delta.c b/diff-delta.c
index 35e517d..c618875 100644
--- a/diff-delta.c
+++ b/diff-delta.c
@@ -136,11 +136,12 @@ struct delta_index {
 
 struct delta_index * create_delta_index(const void *buf, unsigned long bufsize)
 {
-	unsigned int i, hsize, hmask, entries, *hash_count;
+	unsigned int i, hsize, hmask, entries, prev_val, *hash_count;
 	const unsigned char *data, *buffer = buf;
 	struct delta_index *index;
 	struct index_entry *entry, **hash;
 	void *mem;
+	unsigned long memsize;
 
 	if (!buf || !bufsize)
 		return NULL;
@@ -155,9 +156,10 @@ struct delta_index * create_delta_index(
 	hmask = hsize - 1;
 
 	/* allocate lookup index */
-	mem = malloc(sizeof(*index) +
-		     sizeof(*hash) * hsize +
-		     sizeof(*entry) * entries);
+	memsize = sizeof(*index) +
+		  sizeof(*hash) * hsize +
+		  sizeof(*entry) * entries;
+	mem = malloc(memsize);
 	if (!mem)
 		return NULL;
 	index = mem;
@@ -179,18 +181,26 @@ struct delta_index * create_delta_index(
 	}
 
 	/* then populate the index */
-	data = buffer + entries * RABIN_WINDOW - RABIN_WINDOW;
-	while (data >= buffer) {
+	prev_val = ~0;
+	for (data = buffer + entries * RABIN_WINDOW - RABIN_WINDOW;
+	     data >= buffer;
+	     data -= RABIN_WINDOW) {
 		unsigned int val = 0;
 		for (i = 1; i <= RABIN_WINDOW; i++)
 			val = ((val << 8) | data[i]) ^ T[val >> RABIN_SHIFT];
-		i = val & hmask;
-		entry->ptr = data + RABIN_WINDOW;
-		entry->val = val;
-		entry->next = hash[i];
-		hash[i] = entry++;
-		hash_count[i]++;
-		data -= RABIN_WINDOW;
+		if (val == prev_val) {
+			/* keep the lowest of consecutive identical blocks */
+			entry[-1].ptr = data + RABIN_WINDOW;
+		} else {
+			prev_val = val;
+			i = val & hmask;
+			entry->ptr = data + RABIN_WINDOW;
+			entry->val = val;
+			entry->next = hash[i];
+			hash[i] = entry++;
+			hash_count[i]++;
+			entries--;
+		}
 	}
 
 	/*
@@ -220,6 +230,10 @@ struct delta_index * create_delta_index(
 	}
 	free(hash_count);
 
+	/* If we didn't use all hash entries, free the unused memory. */
+	if (entries)
+		index = realloc(index, memsize - entries * sizeof(*entry));
+
 	return index;
 }

Nicolas Pitre· May 3, 2006, 03:46 UTC · re: Nicolas Pitre · lore

[PATCH] tiny optimization to diff-delta

This is my assembly freak side looking at generated code again. And since create_delta() is certainly pretty high on the radar every bits count. In this case shorter code is generated if hash_mask is not copied to a local variable.

Signed-off-by: Nicolas Pitre <nico@cam.org>

---

Show changes to diff-delta.c +2 −3

diff --git a/diff-delta.c b/diff-delta.c
index 26540d8..c618875 100644
--- a/diff-delta.c
+++ b/diff-delta.c
@@ -253,7 +253,7 @@ create_delta(const struct delta_index *i
 	     const void *trg_buf, unsigned long trg_size,
 	     unsigned long *delta_size, unsigned long max_size)
 {
-	unsigned int i, outpos, outsize, hash_mask, val;
+	unsigned int i, outpos, outsize, val;
 	int inscnt;
 	const unsigned char *ref_data, *ref_top, *data, *top;
 	unsigned char *out;
@@ -289,7 +289,6 @@ create_delta(const struct delta_index *i
 	ref_top = ref_data + index->src_size;
 	data = trg_buf;
 	top = trg_buf + trg_size;
-	hash_mask = index->hash_mask;
 
 	outpos++;
 	val = 0;
@@ -304,7 +303,7 @@ create_delta(const struct delta_index *i
 		struct index_entry *entry;
 		val ^= U[data[-RABIN_WINDOW]];
 		val = ((val << 8) | *data) ^ T[val >> RABIN_SHIFT];
-		i = val & hash_mask;
+		i = val & index->hash_mask;
 		for (entry = index->hash[i]; entry; entry = entry->next) {
 			const unsigned char *ref = entry->ptr;
 			const unsigned char *src = data;

patchimprove diff-delta with sparse and/or repetitive data