On Tue, 9 May 2006, Manuel Bouyer wrote:

> On Tue, May 09, 2006 at 11:36:51PM +0200, Tonnerre LOMBARD wrote:
[snip]
>> vic# dd if=/dev/zero of=/dev/rwd0b bs=64k count=1536
>> 100663296 bytes transferred in 41.410 secs (2430893 bytes/sec)
>
> That's way to slow. So it looks like the issue is really somewhere
> between the disk and the controller.
> I would blame the controller, or more probably the way we use it (but without
> doc it's hard to fix this :/

On my previous amd64 box (athlon64 3500+, MSI K8N Neo4 Platinum (Nforce4 
ultra chipset), Seagate barracuda SATA disks) running 3.0 I saw great 
variations in disk speed with dd, depending on block size.

Utilizing ram was a bad idea: 64M blocksize yielded me something like your 
figures (~28MB/s). Too small (8-16K) of course had a similar effect. I got 
a peak rate of 59M/s using a block size of 512K. I got about the same 
results when dd:ing one disk to another and when wiping the old ones with 
dd from /dev/zero. Not sure what I *should* be able to squeeze out of it 
in linear raw-device writes though -- I somehow expected more from a new 
sata disk.

Awaiting the arrical of my new box (x2 4400+, nforce4 sli x16 chipset, 
sata2 barracudas). Will be interesting to see if performance differs.

Regards,
ali:)