[lkml]   [2002]   [Jan]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: hashed waitqueues
On Wed, Jan 16, 2002 at 02:21:41PM +0000, Jamie Lokier wrote:
> Looking up "good hash function" on Google leads to these notable pages:
[various URL's]
> The last one is interesting because it mentions the golden prime
> multiplier function, and suggests good non-multipler functions instead.
> (Justification: the multiplier function doesn't distribute bits evenly).

Excellent! I can always use more of these to test.

It seems odd that they don't like Fibonacci hashing, it appears to pass
various chi^2 tests on bucket distribution. And operator-sparse Fibonacci
hashing primes appear to pass it as well, at least once 10 terms of the
continued fraction match (operator-sparse Fibonacci hashing primes means
that the multiplication can be done with shifts and adds or subtracts).

Regardless, various arches want non-multiplicative hash functions and
they'll be getting them. These hash functions will certainly prove
useful in getting a broader base to test against. I don't care to have
a "pet" hash function, only one that is good as possible.

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2005-03-22 13:23    [W:0.102 / U:10.488 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site