Comparaison between ext4 filesystem with and wihtout i_version using FFSB

Test bed:

Kernel: 2.6.23-ext4-1 and i_version patches posted on 25 oct 2007

Partition size: 70GB 7200rpm disk
IO scheduler: CFQ

Tests are done on a bi-Xeon machine with 2G of RAM and with hyper-threading enabled (-> 4 CPUs).

processor      : 4
vendor_id      : GenuineIntel
cpu family     : 15
model          : 4
model name     : Intel(R) Xeon(TM) CPU 2.80GHz
cpu MHz        : 2793.078
cache size     : 1024 KB
bogomips       : 5586.59

Synopsis:

FFSB filesystem benchmarking software 5.2.1 available at

http://sourceforge.net/projects/ffsb/

FFSB is a multi-threaded filesystem performance measurement tool. For our tests, we define the following profile in ordre to test the impact of the i_version option on a ext4 filesystem.
The create operation is defined, but the major weight is given to the write and append operations, cases where the i_version should be updated quite often.

FFSB profile:

num_filesystems=1
num_threadgroups=1
directio=0
time=1800

[filesystem0]
        location=/mnt/test/
        num_files=20
        num_dirs=4
        max_filesize=33554432
        min_filesize=16777216
[end0]

[threadgroup0]
        num_threads=64
        write_blocksize=65536
        create_weight=1
        write_weight=10
        append_weight=10
[end0]

The partition used is formated with mkfs and mounted with appropriate options before each run.
Filesystems and options are:

	filesystems	mkfs options	mount options
1	ext4 without i_version		extents
2	ext4 with i_version	-I 256	extents,i_version
3	ext3		default option

Results

The following table summarizes the results for each filesystem. The numbers represent the average value of 5 tests.
It seems there are no significant difference enabling the i_version flag or not.

	Ext4 without i_version	Ext4 with i_version	Ext3
Transactions per second	497.68	490.48	458.44
Percent CPU utilization	22.3%	21.9%	19.3%