Sun, 18 Nov 2001 14:55:17 +0000 (UTC), Marcin 'Qrczak' Kowalczyk <qrczak@knm.org.pl> pisze:

> The initial_size argument of Hashtbl.create has much bigger impact on
> efficiency than I thought. The documentation is misleading in saying
> that "The table grows as needed, so n is just an initial guess",
> thus suggesting that it shouldn't matter much what was given after
> the table is filled.

Ah, I see that the version in cvs has changed the policy of resizing
and it indeed works better when a good initial size is not known.
It has a bug though:

--- hashtbl.ml~	Sun Nov 18 17:44:57 2001
+++ hashtbl.ml	Sun Nov 18 17:57:42 2001
@@ -197,7 +197,7 @@
       let bucket = Cons(key, info, h.data.(i)) in
       h.data.(i) <- bucket;
       h.size <- succ h.size;
-      if h.size > Array.length h.data lsl 1 then resize hash h
+      if h.size > Array.length h.data lsl 1 then resize H.hash h
 
     let remove h key =
       let rec remove_bucket = function
@@ -255,7 +255,7 @@
       with Not_found ->
         h.data.(i) <- Cons(key, info, l);
         h.size <- succ h.size;
-        if h.size > Array.length h.data lsl 1 then resize hash h
+        if h.size > Array.length h.data lsl 1 then resize H.hash h
 
     let mem h key =
       let rec mem_in_bucket = function

-- 
 __("<  Marcin Kowalczyk * qrczak@knm.org.pl http://qrczak.ids.net.pl/
 \__/
  ^^
QRCZAK

-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr