[COMMITTED] ada: Tune hash function for cross-reference entries
Checks
Commit Message
From: Piotr Trojanek <trojanek@adacore.com>
Tune the hash function that combines entity identifiers with source
locations of where those entities are referenced. Previously the source
location was multiplied by 2 ** 7 (i.e. shifted left by 7 bits), then
added to the entity identifier, and finally divided modulo 2 ** 16 (i.e.
masked to only use the lowest 16 bits). This hash routine caused
collisions that could make some tests up to twice slower.
With a large entity number the source location was only contributing few
bits to the hash value. This large entity number might correspond to
entity like Ada.Characters.Latin_1.NUL that occurs thousands of times in
generated code.
gcc/ada/
* lib-xref.adb (Hash): Tune hash function.
Tested on x86_64-pc-linux-gnu, committed on master.
---
gcc/ada/lib-xref.adb | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
@@ -1271,10 +1271,10 @@ package body Lib.Xref is
XE : Xref_Entry renames Xrefs.Table (F);
type M is mod 2**32;
- H : constant M := M (XE.Key.Ent) + 2 ** 7 * M (abs XE.Key.Loc);
+ H : constant M := 3 * M (XE.Key.Ent) + 5 * M (abs XE.Key.Loc);
-- It would be more natural to write:
--
- -- H : constant M := M'Mod (XE.Key.Ent) + 2**7 * M'Mod (XE.Key.Loc);
+ -- H : constant M := 3 * M'Mod (XE.Key.Ent) + 5 * M'Mod (XE.Key.Loc);
--
-- But we can't use M'Mod, because it prevents bootstrapping with older
-- compilers. Loc can be negative, so we do "abs" before converting.