Skip to content

Commit 38daaea

Browse files
authored
README: Add note about Concurrent Hash Table vs this approach
1 parent dbb28a2 commit 38daaea

File tree

1 file changed

+4
-0
lines changed

1 file changed

+4
-0
lines changed

README.md

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -18,6 +18,10 @@ you might use instead.
1818
## Alternatives Considered
1919
Some other approaches to concurrent caches include:
2020
- Concurrent Hash Table
21+
- The _theory_ in this package is to take advantage of the append-only aspect of a Cache
22+
to get some contention benefits over a conventional multi-threaded dictionary, but we're
23+
less sure in practice if this is actually true. (Once we had to make the adjustments to
24+
account for task migration, the benefits of the current design became less clear...)
2125
- Multithread caches designed for low contention by sharding the **key space**, and keeping
2226
a separate lock per sub-cache. For example, having an array of e.g. 64 separate caches,
2327
sharded by a prefix of a key's hash, each with their own lock.

0 commit comments

Comments
 (0)