Barracuda XT, Barracuda, Barracuda Green, DiamondMax and other desktop SATA and IDE drives

[Kristopher]

I am posting an e-mail received from HD Tune support with regards to the recent postings regarding 0MB cache sizes being reported on 7200.11 drives with AD14 firmware.  After speaking with Seagate technical support & upgrading my firmware, I was having the same issue and decided to contact a third-party for some technical insight.  I did this because I wanted to attempt validate Seagate's response that the cache was there but could not be seen.  It sounds like what the Seagate support said was valid unless some other information comes to light.

========================================
From: HD Tune support [support@hdtune.com]
Sent: 11 February 2008 13:09
To: xxx, Kristopher
Subject: Re: ST3750330AS Reporting 0MB Cache...

Hello,

Because of a limitation of the ATA specification, buffer lengths of 32 MB and higher cannot be reported by the hard disk, so they will return '0'
instead.
To avoid confusion the next version of HD Tune wil show 'n/a' instead of '0 MB' for hard disks with buffer lengths of 32MB and above.

Best regards
------------------------------------------------------
HD Tune
http://www.hdtune.com
------------------------------------------------------

----- Original Message -----
From: <Kristopher.xxx@xxx.com>
To: <support@hdtune.com>
Sent: Monday, February 11, 2008 1:15 PM
Subject: ST3750330AS Reporting 0MB Cache...

I have noticed a lot of e-mails on various groups, including Seagates
own support website with regards to the Barracuda 7200.11 hard drives
ST3500320AS, ST3750330AS, & ST31000340AS.  The reason I am contacting
you is because the drive is reporting 0MB of cache and Seagate Level III
support has stated that their drives do infact have the cache but HDTune
is incorrectly reporting this amount.  I cannot fathom why this would be
unless they have somehow changed their protocol reporting the cache size
or unless it is due to the fact that drives with 32MB only came out
recently.  I was just wanting to be able to verify their claim or have
some knowledge to go back to them with to get the drive replaced.  As a
side note, othe software, such as S.A.N.D.R.A., is also reporting 0MB or
cache not present.

regards,
-Kristopher

[Chuck]

Kristopher,

HD Tune reports the same thing for my drive with fw SD04. Yet the fw upgrade does not apply to the model I have. I have ST3500320AS which does not need a fw upgrade apparently. So I ama alittle confused about this whole fw and model number business. Do I need to upgrade the firmware or not?

[Kristopher]

Per the Seagate support web page regarding drives with SD04 or SD14 firmware, there is no need to upgrade the firmware if your drive is displaying the model number correctly.  Quoting from the support page, 'A small number of Barracuda 7200.11 drives with firmware versions SD04 or SD14 may show and utilize only part of the available cache.'  Therefore, I assume that the majority of the drives with firmware SD04 or SD14 are not affected and need no update.  The easiest way is to check in your BIOS - if the model number displayed does not match the model number of the drive, you would need to apply the firmware upgrade.  In my case, the BIOS was displaying ST3_INVALID_PFM when it should have been displaying ST3750330AS.  After upgrading my firmware to AD14, the model number is now correctly displayed in the BIOS.  In your case, your model number is displayed correctly, so you should not need to apply the BIOS update.  If in doubt, contact Seagate technical support for further clarification.

regards,
-Kristopher

[Chuck]

Kristopher,

It done and my drive has been flashed. It worked because the models listed in the flash program recognized my drive. Did not seem to make much difference before or after.

[NeGrusti]

Well...
Long time ago, in the specification there were 2 bytes (a word) reserved for cache size in sectors. As you can easily calculate, 32Mb cache size in sectors is 0x10000 an it cannot fit into two bytes, what leaves us with 0x0000 = 0 value...

[noegruts]

I tried to find exactly where in the ATA spec. the command to obtain the cache size is described, but I couldn't find it.

Your speculation sounds correct, but I can't confirm it.  Can you point to a spec. that defines this aspect of the ATA command set?

[NeGrusti]

No speculations, just pure knowledge
I said "long time ago":
ftp://ftp.t13.org/project/d0791r4c-ATA-1.pdf
page 44
or
http://seagate.custhelp.com/cgi-bin/seagate.cfg/php/enduser/std_adp.php?p_faqid=5224

Yes, it's been deprecated AND actively used THAT long time


(Edited long URL into shorter, still-functional URL)


Message Edited by AlanM on 03-25-2008 07:37 AM

[emdeesee]

So I guess the "million dollar question" is:

Ho do we KNOW that the full 32MB of onboard cache is being utilized? I mean seriously.. I could care less if benchmark and SMART-reading programs can't report it (NOT Seagate's fault).. but the 32MB cache is what I bought this HD for.. over all others (that.. and it's only two platters) and I'd LIKE to know that it's at least being seen and utilized by my OS (WXP Pro).

Has anybody found any program yet that sees this?

Also.. the new ST3500320AS drive I just recieved and installed two days ago has firmware rev. AD14 on it (as per label). Although "A" comes before "S".. it took a bit of research for me to deduce that AD14 is indeed the last firmware; and HD Tune pro (v3.00) does indeed now report my buffer as "n/a".

[NeGrusti]

OS has no access to the hard drive buffer, only hard drive firmware does. So OS counldn't care less... No software rely on this parameter for it's operation. If you want to satisfy you curiosity an verify there is indeed 32Mb cache on the drive, read markings on the RAM chip on the electronics board and find a datasheet online.