lkml.org 
[lkml]   [2007]   [May]   [29]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: [RFC] LZO de/compression support - take 6
Date
All problems I was having with the test-bed code have been solved, and the 
error I was running into was, as I suspected, in the code I used to fill the
buffer for the random-data test.

Results of running the new benchmark (version 6 of the benchmark, version 6
of 'tinyLZO'):
10000 run averages:
'Tiny LZO':
Combined: 104.1529 usec
Compression: 43.065 usec
Decompression: 18.4648 usec
Random Data Compression: 31.132 usec
Random Data Decompression: 11.4911 usec
'miniLZO':
Combined: 106.1363 usec
Compression: 44.6988 usec
Decompression: 18.3576 usec
Random Data Compression: 31.5772 usec
Random Data Decompression: 11.5027 usec
Percentages (calculated as: ((full - tiny)/full)*100):
Overall (combined totals): Tiny is 1.87 % faster
Compression: Tiny is 3.66 % faster
Decompression: Tiny is 0.58 % slower
Random Compression: Tiny is 1.41 % faster
Random Decompression: Tiny is 0.10 % faster

The results, on my system, are not consistent, except in that 'TinyLZO' is
generally faster in the non-random data tests than miniLZO. It did, three of
the five times I ran the test, perform (on average) about 1% worse than
miniLZO.

In order to sidestep any issues that a difference in the input data might have
caused, I'm going to rewrite the code so that all tests are run against the
same data. However, in the meantime, I've attached the latest version of my
test-code.

DRH
PS: the code is going to massively change as I work to include more data
sources for the benchmarking, as well as tests that will try to really stress
the stripped-down code.
[unhandled content-type:application/x-tbz]
\
 
 \ /
  Last update: 2007-05-29 23:51    [W:0.268 / U:0.104 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site