lkml.org 
[lkml]   [1998]   [Aug]   [7]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: 2.1.114 VFS code 5x slower than 2.0.33?!?
From
Date
Zlatko Calusic <Zlatko.Calusic@CARNet.hr> writes:

> Anybody care to shed some light on this?

Not really light, but

I repeated your tests with kernel profiling on 2.1.114:

Creation:
28.42user 59.63system 2:18.66elapsed 63%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (710242major+390114minor)pagefaults 0swaps

...
64 1.05% c011ab7c zap_page_range
68 1.11% c011a9b4 copy_page_range
81 1.33% c01383e0 ext2_get_group_desc
95 1.56% c0126134 get_hash_table
98 1.61% c010ada0 show_registers (this appears often in
other kernel profiles too - anyone know the reason?)
111 1.82% c013a994 ext2_new_inode
192 3.16% c011d988 filemap_nopage
382 6.28% c01396c0 ext2_readdir
463 7.62% c011b384 do_wp_page
823 13.54% c013c6e0 ext2_find_entry
953 15.68% c013ca38 ext2_add_entry
1263 20.79% c013960c ext2_check_dir_entry
6074 100.00% 00000000 total
List:

0.57user 10.42system 0:11.16elapsed 98%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (123major+512minor)pagefaults 0swaps
8651
12 1.13% c013b9a0 ext2_getblk
13 1.23% c0126814 getblk
30 2.84% c0126134 get_hash_table
92 8.72% c01396c0 ext2_readdir
354 33.58% c013960c ext2_check_dir_entry
487 46.20% c013c6e0 ext2_find_entry
1054 100.00% 00000000 total
Deletion:

0.78user 15.06system 0:44.03elapsed 35%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (81major+20minor)pagefaults 0swaps
16 1.04% c010ab28 system_call
17 1.11% c012c0bc lookup_dentry
19 1.24% c013082c d_lookup
21 1.37% c0126b80 refile_buffer
25 1.63% c013b9a0 ext2_getblk
32 2.09% c0126814 getblk
62 4.06% c0126134 get_hash_table
122 7.98% c01396c0 ext2_readdir
411 26.91% c013960c ext2_check_dir_entry
563 36.86% c013c6e0 ext2_find_entry
1527 100.00% 00000000 total
All running on a 128MB PII-300 machine with 2.1.114/UP and a 2940UW
controller. I resetted the profile counters before each run, and took
the statistics after the disk thrashing stopped.

One possible reason for the slowdown on your side: if you're using
IDE note that DMA defaults to off now, which slows down a lot of things.
I did my tests with a SCSI system.

-Andi

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html

\
 
 \ /
  Last update: 2005-03-22 13:43    [from the cache]
©2003-2011 Jasper Spaans