comment

doc/bugs/importtree_spends_hours_reading_cidsdb/comment_6_b8b3eb5fff4ad3478dfbbd2120d4b00a._comment

@@ -0,0 +1,35 @@
+[[!comment format=mdwn
+ username="joey"
+ subject="""Results with 150000 files"""
+ date="2023-06-01T18:17:34Z"
+ content="""
+Initial sync: 39:01.37
+
+Then `git-annex import master:testdata --from source` 
+with no changed files: 8:36.30  
+(This is writing the cidsdb since the initial sync no 
+longer does, see comment #2)
+
+Subsequent imports with no changed files: 0:37.21
+
+Subsequent syncs with no changed files: 3:18.61
+
+Subsequent syncs with no changed files, when not on an adjusted branch: 0:38.40
+
+That 37 second run time for an import with no changed files seems to
+indicate that my optimisation so far was successful at eliminating all
+cidsdb reading. It's bottlenecked on git operations now.
+
+I also tried an initial import of 150000 files to see how much of that
+initial sync was other work than the import (updating adjusted branch,
+checking out files, etc).
+
+Initial import: 38:24.36
+
+So the initial sync is mostly spending time on imports, not other stuff.
+It could perhaps be sped up some. Of course it does support -J2 which may
+speed it up throuh parallism.
+
+Seems like a sync with no changed files on an adjusted branch is doing
+unncessary work to update the adjusted branch when nothing has changed.
+"""]]