Allow the bloom size to vary depending on how many fetchable keys there
are - so ther eis no large bloom held if most of the keys are index
entries for example
the Skiplist range code was needlessly complicated. It may be faster
than the new code, but the complexity delta cannot be support for such a
small change.
This was incovered whilst troubleshooting the initial kv range test.
Changing the slot size higher has a significant impact on the fetch
time, although it allows for more constant time on write. i.e. doubling
the size means 5 x cost of read, if only a 10% increase at write time.
Move to using the DJ Bernstein Magic Hash consistently, and trying to
make sure we only hash once for each operation (as the hash is more
expensive than phash2).
The improved lookup time for missing keys should allow for the L0 index
to be removed, and hence speed up the completion time for push_mem
operations.
It is expected there will be a second stage of creating a tinybloom as
part of the SFT creation process, and then adding that tinybloom to the
manifest. This will then reduce the message passing required for a GET
not in the cache or higher levels
