lkml.org 
[lkml]   [1996]   [Mar]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectFilesystem performance in 1.3.78
Date
Here are some results from Bonnie with kernels 1.3.77 and 1.3.78. The results
are medium values from three tests. Sequential input per block has increased
a little bit, as it perhaps should with Linus' new code in ll_rw_block.c, but
what happened to sequential input per char? Shouldn't CPU utilization be
closer to 100%. I also tried to change the read_ahead value, but no visible
difference could be found.
Another question: Why is the random seeks results better with 1.3.78 ??

-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
1.3.77 70 3172 72.1 3266 20.5 1403 20.1 3490 82.1 4143 26.3 57.2 3.0

-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
1.3.78 70 3169 71.2 3269 19.3 1428 17.5 2961 63.7 4303 24.3 68.8 3.7

My system is a 100Mhz Pentium with 16MB RAM, NCR810 and a Seagate ST31230N
(1 GB Hawk 2LP). The filesystem had only 100 MB free, hence the small filesize.
Now is the time to buy another drive and stripe them!

With best regards from Olov want-to-have-the-most-out-of-his-Linux-box Karlsson

Olov Karlsson dvlokn@cs.umu.se http://www.ts.umu.se/~olov



\
 
 \ /
  Last update: 2005-03-22 13:36    [W:0.036 / U:0.212 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site