git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH v3 8/8] diffcore-rename: avoid doing basename comparisons for irrelevant sources

From
Elijah Newren via GitGitGadget <gitgitgadget@gmail.com>
Date
Mar 11, 2021, 00:38 UTC
Message-ID
<98b0c7de5e70d62d47c3eeb3d290c6a234214f40.1615423112.git.gitgitgadget@gmail.com>
In-Reply-To
<pull.845.v3.git.1615423111.gitgitgadget@gmail.com>
From: Elijah Newren <newren@gmail.com>

The basename comparison optimization implemented in find_basename_matches() is very beneficial since it allows a source to sometimes only be compared with one other file instead of N other files. When a match is found, both a source and destination can be removed from the matrix of inexact rename comparisons. In contrast, the irrelevant source optimization only allows us to remove a source from the matrix of inexact rename comparisons...but it has the advantage of allowing a source file to not even be loaded into memory at all and be compared to 0 other files. Generally, not even comparing is a bigger performance win, so when both optimizations could apply, prefer to use the irrelevant-source optimization.

For the testcases mentioned in commit 557ac0350d ("merge-ort: begin performance work; instrument with trace2_region_* calls", 2020-10-28), this change improves the performance as follows:

                            Before                  After
    no-renames:        5.708 s ±  0.111 s     5.680 s ±  0.096 s
    mega-renames:    102.171 s ±  0.440 s    13.812 s ±  0.162 s
    just-one-mega:     3.471 s ±  0.015 s   506.0  ms ±  3.9  ms
Signed-off-by: Elijah Newren <newren@gmail.com>
---
 diffcore-rename.c | 37 +++++++++++++++++++++++++++++++++----
 1 file changed, 33 insertions(+), 4 deletions(-)
diff --git a/diffcore-rename.c b/diffcore-rename.c
index 7f6115fd9018..e8508541be14 100644
--- a/diffcore-rename.c
+++ b/diffcore-rename.c
@@ -527,6 +527,7 @@ static void update_dir_rename_counts(struct dir_rename_info *info,
 }
 
 static void initialize_dir_rename_info(struct dir_rename_info *info,
+				       struct strset *relevant_sources,
 				       struct strset *dirs_removed,
 				       struct strmap *dir_rename_count)
 {
@@ -534,7 +535,7 @@ static void initialize_dir_rename_info(struct dir_rename_info *info,
 	struct strmap_entry *entry;
 	int i;
 
-	if (!dirs_removed) {
+	if (!dirs_removed && !relevant_sources) {
 		info->setup = 0;
 		return;
 	}
@@ -549,7 +550,20 @@ static void initialize_dir_rename_info(struct dir_rename_info *info,
 	strmap_init_with_options(&info->dir_rename_guess, NULL, 0);
 
 	/* Setup info->relevant_source_dirs */
-	info->relevant_source_dirs = dirs_removed;
+	info->relevant_source_dirs = NULL;
+	if (dirs_removed || !relevant_sources) {
+		info->relevant_source_dirs = dirs_removed; /* might be NULL */
+	} else {
+		info->relevant_source_dirs = xmalloc(sizeof(struct strintmap));
+		strset_init(info->relevant_source_dirs);
+		strset_for_each_entry(relevant_sources, &iter, entry) {
+			char *dirname = get_dirname(entry->key);
+			if (!dirs_removed ||
+			    strset_contains(dirs_removed, dirname))
+				strset_add(info->relevant_source_dirs, dirname);
+			free(dirname);
+		}
+	}
 
 	/*
 	 * Loop setting up both info->idx_map, and doing setup of
@@ -627,6 +641,13 @@ static void cleanup_dir_rename_info(struct dir_rename_info *info,
 	/* dir_rename_guess */
 	strmap_clear(&info->dir_rename_guess, 1);
 
+	/* relevant_source_dirs */
+	if (info->relevant_source_dirs &&
+	    info->relevant_source_dirs != dirs_removed) {
+		strset_clear(info->relevant_source_dirs);
+		FREE_AND_NULL(info->relevant_source_dirs);
+	}
+
 	/* dir_rename_count */
 	if (!keep_dir_rename_count) {
 		partial_clear_dir_rename_count(info->dir_rename_count);
@@ -749,6 +770,7 @@ static int idx_possible_rename(char *filename, struct dir_rename_info *info)
 static int find_basename_matches(struct diff_options *options,
 				 int minimum_score,
 				 struct dir_rename_info *info,
+				 struct strset *relevant_sources,
 				 struct strset *dirs_removed)
 {
 	/*
@@ -839,6 +861,11 @@ static int find_basename_matches(struct diff_options *options,
 		intptr_t src_index;
 		intptr_t dst_index;
 
+		/* Skip irrelevant sources */
+		if (relevant_sources &&
+		    !strset_contains(relevant_sources, filename))
+			continue;
+
 		/*
 		 * If the basename is unique among remaining sources, then
 		 * src_index will equal 'i' and we can attempt to match it
@@ -1164,7 +1191,7 @@ void diffcore_rename_extended(struct diff_options *options,
 
 		/* Preparation for basename-driven matching. */
 		trace2_region_enter("diff", "dir rename setup", options->repo);
-		initialize_dir_rename_info(&info,
+		initialize_dir_rename_info(&info, relevant_sources,
 					   dirs_removed, dir_rename_count);
 		trace2_region_leave("diff", "dir rename setup", options->repo);
 
@@ -1172,7 +1199,9 @@ void diffcore_rename_extended(struct diff_options *options,
 		trace2_region_enter("diff", "basename matches", options->repo);
 		rename_count += find_basename_matches(options,
 						      min_basename_score,
-						      &info, dirs_removed);
+						      &info,
+						      relevant_sources,
+						      dirs_removed);
 		trace2_region_leave("diff", "basename matches", options->repo);
 
 		/*
-- 
gitgitgadget
Previous: Elijah Newren via GitGitGadgetNext: Elijah Newren via GitGitGadget
Message 35 of 39 in “Optimization batch 9: avoid detecting irrelevant renames”
  1. 0/8 Optimization batch 9: avoid detecting irrelevant renamesElijah Newren via GitGitGadget, Feb 28, 2021
  2. 1/8 diffcore-rename: enable filtering possible rename sourcesElijah Newren via GitGitGadget, Feb 28, 2021
  3. 3/8 merge-ort: add data structures for an alternate tree traversalElijah Newren via GitGitGadget, Feb 28, 2021
  4. 2/8 merge-ort: precompute subset of sources for which we need rename detectionElijah Newren via GitGitGadget, Feb 28, 2021
  5. 4/8 merge-ort: introduce wrappers for alternate tree traversalElijah Newren via GitGitGadget, Feb 28, 2021
  6. 5/8 merge-ort: precompute whether directory rename detection is neededElijah Newren via GitGitGadget, Feb 28, 2021
  7. 6/8 merge-ort: use relevant_sources to filter possible rename sourcesElijah Newren via GitGitGadget, Feb 28, 2021
  8. 7/8 merge-ort: skip rename detection entirely if possibleElijah Newren via GitGitGadget, Feb 28, 2021
  9. 8/8 diffcore-rename: avoid doing basename comparisons for irrelevant sourcesElijah Newren via GitGitGadget, Feb 28, 2021
  10. Elijah NewrenMar 1, 2021
  11. Elijah NewrenMar 4, 2021
  12. 0/8 Optimization batch 9: avoid detecting irrelevant renamesElijah Newren via GitGitGadget, Mar 9, 2021
  13. 2/8 merge-ort: precompute subset of sources for which we need rename detectionElijah Newren via GitGitGadget, Mar 9, 2021
  14. 1/8 diffcore-rename: enable filtering possible rename sourcesElijah Newren via GitGitGadget, Mar 9, 2021
  15. Derrick StoleeMar 9, 2021
  16. Elijah NewrenMar 9, 2021
  17. Derrick StoleeMar 9, 2021
  18. 6/8 merge-ort: use relevant_sources to filter possible rename sourcesElijah Newren via GitGitGadget, Mar 9, 2021
  19. 3/8 merge-ort: add data structures for an alternate tree traversalElijah Newren via GitGitGadget, Mar 9, 2021
  20. 4/8 merge-ort: introduce wrappers for alternate tree traversalElijah Newren via GitGitGadget, Mar 9, 2021
  21. Derrick StoleeMar 9, 2021
  22. 5/8 merge-ort: precompute whether directory rename detection is neededElijah Newren via GitGitGadget, Mar 9, 2021
  23. 8/8 diffcore-rename: avoid doing basename comparisons for irrelevant sourcesElijah Newren via GitGitGadget, Mar 9, 2021
  24. 7/8 merge-ort: skip rename detection entirely if possibleElijah Newren via GitGitGadget, Mar 9, 2021
  25. Derrick StoleeMar 9, 2021
  26. Elijah NewrenMar 9, 2021
  27. Derrick StoleeMar 9, 2021
  28. Derrick StoleeMar 10, 2021
  29. 0/8 Optimization batch 9: avoid detecting irrelevant renamesElijah Newren via GitGitGadget, Mar 11, 2021
  30. 1/8 diffcore-rename: enable filtering possible rename sourcesElijah Newren via GitGitGadget, Mar 11, 2021
  31. 3/8 merge-ort: add data structures for an alternate tree traversalElijah Newren via GitGitGadget, Mar 11, 2021
  32. 5/8 merge-ort: precompute whether directory rename detection is neededElijah Newren via GitGitGadget, Mar 11, 2021
  33. 2/8 merge-ort: precompute subset of sources for which we need rename detectionElijah Newren via GitGitGadget, Mar 11, 2021
  34. 4/8 merge-ort: introduce wrappers for alternate tree traversalElijah Newren via GitGitGadget, Mar 11, 2021
  35. 8/8 diffcore-rename: avoid doing basename comparisons for irrelevant sourcesElijah Newren via GitGitGadget, Mar 11, 2021
  36. 6/8 merge-ort: use relevant_sources to filter possible rename sourcesElijah Newren via GitGitGadget, Mar 11, 2021
  37. 7/8 merge-ort: skip rename detection entirely if possibleElijah Newren via GitGitGadget, Mar 11, 2021
  38. Derrick StoleeMar 15, 2021
  39. Elijah NewrenMar 15, 2021

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.