ciss(4): speed degradation for Compaq Smart Array

[Maxim Sobolev]

Are you sure that in both cases /tmp is located in the same physical 
zone of the disk? Due to CAV (constant angular velocity) natute of 
modern HDDs there is drastical difference (more than 2x) in read/write 
speeds between data located on outer tracks and data on inner tracks. 
Therefore, to make a meaningful comparison you should at least mount the 
very same slice in every case. Even better, you need to newfs this slice 
before doing any tests, to avoid any interference of your results with 
fragmentation.

-Maxim

Andrey Koklin wrote:
> We have several Proliants DL380 with integrated Smart Array 5i controller,
> with installed FreeBSD OS versions 4.11-stable, 5.4-pre and 6.0-current.
> All systems have disks configured as RAID5.
> 
> I've noticed there is drastic difference in speed for the systems.
> While I have linear transfer 45-50 Mb/s (write) and 100 Mb/s (read)
> for FreeBSD 4.11, there is near two times degradation in speed
> with 5.4 and 6.0.
> 
> -- 8< ------------------------------------------------------------------
> 
> FreeBSD 4.11:
> ciss.c,v 1.2.2.21 2005/01/21
> 
> # dd if=/dev/zero of=/var/tmp/4G bs=1024 count=$[4*1024*1024]
> 
> 4294967296 bytes transferred in 107.075789 secs (40111470 bytes/sec)
> 
> # dd if=/var/tmp/4G of=/dev/zero
> 
> 4294967296 bytes transferred in 42.394322 secs (101309966 bytes/sec)
> 
> 
> FreeBSD 5.4-pre:
> ciss.c,v 1.56.2.1 2005/01/20
> 
> # dd if=/dev/zero of=/var/tmp/4G bs=1024 count=$[4*1024*1024]
> 
> 4294967296 bytes transferred in 163.088095 secs (26335259 bytes/sec)
> 
> # dd if=/var/tmp/4G of=/dev/zero
> 
> 4294967296 bytes transferred in 79.773797 secs (53839324 bytes/sec)
> 
> 
> FreeBSD 6.0:
> ciss.c,v 1.60 2005/03/29
> 
> # dd if=/dev/zero of=/var/tmp/4G bs=1024 count=$[4*1024*1024]
> 
> 4294967296 bytes transferred in 224.804547 secs (19105340 bytes/sec)
> 
> # dd if=/var/tmp/4G of=/dev/zero
> 
> 4294967296 bytes transferred in 94.839699 secs (45286598 bytes/sec)
> 
> -- 8< ------------------------------------------------------------------
> 
> These are simple tests taken on running systems (tests in single user
> mode are slightly different but of the same order)
> 
> It seems, there is sensible speed penalty with driver version growing.
> 
> Andrey
> _______________________________________________
> freebsd-current at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to "freebsd-current-unsubscribe at freebsd.org"
> 
> 
>