List of fake vs. real SATA drives.

Søren Schmidt sos at DeepCore.dk
Mon Nov 22 22:44:26 GMT 2004


Stephen Montgomery-Smith wrote:
> Thomas Wolf wrote:
> 
>> Stephen Montgomery-Smith <stephen at math.missouri.edu> schrieb:
>>
>>> FUJISHIMA Satsuki wrote:
>>>
>>>> Currently native SATA drives are still not so popular. There are:
>>>
>>>> Seagate Barracuda ATA V, 7200.7, 7200.8
>>>
>>> I have one of these, and I am really impressed by its performance.  I 
>>> added one to my computer, which came with a Maxtor 6Y080L0.  My main 
>>> disk intensive operation is creating the CTM deltas (as in CTM which 
>>> is an alternative to CVSUP for people behind unfriendly firewalls).  
>>> The performance difference was somewhat collosal, as in something 
>>> like 3 times faster.  To be honest, I am still at a loss to explain 
>>> why the Seagate did so very much better - maybe it is the 8M cache as 
>>> compared to the 2M cache.  The Seagate 7200.7 had similar performance 
>>> to a Seagate 160MHz SCSI drive that I have on another computer.
>>
>> Ah, please tell me more about it, is this a ST3120827AS?
>> (I would need the exact PartNo.) What controller dou you have and 
>> finally, on which version of FreeBSD?
>>
> ST380013AS with Intel ICH5 SATA150 controller and FreeBSD 4.10-Stable. 
> The kernel reports that it is running at UDMA33, but the actual 
> performance seems much better than that.

Hold on right there, the ST380013AS is a Barracuda V 7200.7 device and 
does not do any form of tagging, neither does the ICH5 support it. You 
are just enjoying the stock speed of modern (S)ATA gear :)
Oh, and 4.x has no notion of SATA it only works on controllers that can 
emulate the old ATA way of things. This is means that the ATA33 speed 
isn't whats used as SATA v1 always runs at 150MB/s, just 4.x has no way 
of telling.

-- 

-Søren




More information about the freebsd-hackers mailing list