Murmurhash is generally superior to the Jenkins lookup3 hash according to
the available figures. Perhaps we should generally replace our current
hashes by murmurhash.
For now, I'm introducing a parallel implementation to allow it to be used
in cases where an incremental one-word-at-a-time hash is desirable. The
first user will be added in an upcoming commit.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Ethan Jackson <ethan@nicira.com>
Replaced all instances of Nicira Networks(, Inc) to Nicira, Inc.
Feature #10593
Signed-off-by: Raju Subramanian <rsubramanian@nicira.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
I always assumed that macros would generate better code. I was wrong. The
generated code was identical with inline functions, with GCC version 4.4.5.
Signed-off-by: Ben Pfaff <blp@nicira.com>
This commit makes several library headers suitable for inclusion in C++.
It adds [extern "C"] guards and makes minor changes to fix casting and
keyword issues.
It's easier for the client to pass in data directly than it is for the
client to stuff it into an array.
These functions will have new users in upcoming commits.
Some versions of GCC complain about violations of strict aliasing
rules due to a cast between different pointer types. This avoids
any aliasing by copying the value first.
As previously defined, the following both returned the same value for
given values of 'basis':
hash_int(0, hash_int(1, basis))
hash_int(1, hash_int(0, basis))
because hash_int(0, basis) evaluated to basis and hash_int(1, basis)
evaluated to c + basis for some constant c.
This commit fixes the problem, by eliminating any simple linear
relationship between basis and the hash value.
We should write some tests for hash function quality.
