Some notes on used key hashing algorithm
git-svn-id: http://repo.ham.fi/svn/aprsc/trunk@210 3ce903b1-3385-4e86-93cd-f9a4a239f7ac
This commit is contained in:
parent
7d0a884e6f
commit
181abd575c
|
|
@ -94,15 +94,12 @@ cfgfile.c
|
|||
A configuration file parser written by Tomi Manninen, OH2BNS,
|
||||
originally for the node(1) program in the ax25-utils package.
|
||||
|
||||
splay.c
|
||||
spsymbol.c
|
||||
splay tree functions written by David Brower, daveb@rtech.uucp 1/88,
|
||||
obtained from the zmailer sources.
|
||||
|
||||
crc32.c
|
||||
CRC32 calculation routine by Rayan S. Zachariassen, originally
|
||||
from the zmailer MTA software (maintained by Matti Aarnio,
|
||||
OH2MQK). Used by spsymbol.c.
|
||||
keyhash.c
|
||||
Various attempts at keyhashing.
|
||||
Contains algorithms:
|
||||
- FNV-1a
|
||||
- CRC32
|
||||
Currently using FNV-1a
|
||||
|
||||
rwlock.c
|
||||
A schoolbook pthread rwlock implementation for systems
|
||||
|
|
|
|||
|
|
@ -19,13 +19,8 @@
|
|||
*
|
||||
*/
|
||||
|
||||
/* ON THIS PROCESSING ALONE THE SYSTEM IS PUSHING AROUND 8-12 % OF CPU TIME!
|
||||
* ... but the previous top waster, the aprsc output filters, are optimized
|
||||
* to the hilt...
|
||||
*
|
||||
* There exists alternate implementations of CRC32 which are 1.7 - 2.3 times
|
||||
* faster than this one with an expense of using 4kB / 8 kB / 16 kB of tables,
|
||||
* which of course fill caches...
|
||||
/*
|
||||
* Keyhash routines for the system.
|
||||
*
|
||||
* What is needed is _fast_ hash function. Preferrably arithmethic one,
|
||||
* which does not need table lookups, and can work with aligned 32 bit
|
||||
|
|
@ -36,6 +31,9 @@
|
|||
* http://www.ibiblio.org/pub/Linux/devel/lang/c/mph-1.2.tar.gz
|
||||
* http://www.concentric.net/~Ttwang/tech/inthash.htm
|
||||
* http://isthe.com/chongo/tech/comp/fnv/
|
||||
*
|
||||
* Currently using FNV-1a
|
||||
*
|
||||
*/
|
||||
|
||||
/*
|
||||
|
|
@ -60,6 +58,19 @@
|
|||
*/
|
||||
|
||||
|
||||
/*
|
||||
* ON THIS CRC32 HASH CALCULATION PROCESSING ALONE THE SYSTEM IS PUSHING
|
||||
* AROUND 8-12 % OF CPU TIME!
|
||||
*
|
||||
* ... but the previous top waster, the aprsc output filters, are optimized
|
||||
* to the hilt...
|
||||
*
|
||||
* There exists alternate implementations of CRC32 which are 1.7 - 2.3 times
|
||||
* faster than this one with an expense of using 4kB / 8 kB / 16 kB of tables,
|
||||
* which of course fill caches...
|
||||
*
|
||||
*/
|
||||
|
||||
/*
|
||||
* Oct 15, 2000 Matt Domsch <Matt_Domsch@dell.com>
|
||||
* Nicer crc32 functions/docs submitted by linux@horizon.com. Thanks!
|
||||
|
|
|
|||
Loading…
Reference in New Issue