The new code isn't necessarily much faster on these cases. But it
uses less memory for writing now that writes go directly to
the page cache instead of through the buffer cache first.
So testing massive writing on a slightly memory-constrained
machine might show a difference.
The buffer cache needs re-tuning now that the burden of
writes is removed from it. This isn't done yet.
Linus suggested some kinds of databases, and I seem to recall
someone posting an impressive database result for 2.3.7.
See the archives.
Helge Hafting
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/