Hello Folks!

This is my first post here. I'm rather more advanced than most folks, but I would hardly say I'm a hard disk drive guru ... although I did write one or two disk drivers a long, long time ago.

I recently purchased a new Seagate Barracuda 7200.11 ST31500341AS 1.5 TB drive with CC1H firmware. The data sheet from Seagate (URL below) shows this drive as having 32 MB cache.

http://www.seagate.com/docs/pdf/datashe ... 200_11.pdf

When I run the cache test in HD Tune Pro v4.60 on the drive I get a (fairly) straight line across the top of the graph at 60 MB/sec. from 1 MB to 64 MB.

This would seem to indicate that the drive has 64 MB of cache.

Is there any other test I can run on the ST31500341AS that would TEST the drive for the amount of cache on the PCB? As a "non-drive-repair-professional" freeware or open source would be best as I'll likely only be using the recommended software once for this specific test. Any Windows / Linux / PC Unix suggestions would be welcome.

On the other hand maybe the PC build shop folks down the road a bit might have some other "normal payware" software that you folks could recommend. I might be able to swap something with them for such a small test.

I've spent multiple hours poking around the web and this website for clues about where to find such cache test software. If I've missed a posting on this website that already has this info then please point me to it.

Thanx.

You're a guru.


The second "3" in the model number indicates that the drive has 32MB cache.


I believe Seagate's SeaTools may be able to detect the cache size using vendor specific ATA commands.

Alternatively, since you only need to do this once, remove the PCB from the drive and record the markings on the SDRAM chip. Then refer to its datasheet.

See my notes:
http://www.users.on.net/~fzabkar/HDD/HDD_ICs.txt

As for HD Tune Pro's cache test, could you show us the graph?

AIUI, the transfer rate out of cache should be approaching the interface speed, ie 300MB/s. In fact the drive's maximum sustained data rate off the platters is 130 MB/s, so the data rate into and out of cache must be much faster than that. ISTM that your system must have a bottleneck somewhere.

_________________
A backup a day keeps DR away.

Yes. The data sheet for the drive indicates it should have 32 MB cache.



I ran SeaTools v2.22 (GUI) for DOS and the "short test" said the drive was "OK." There was no option to get further information. From the SeaTools for Windows User Guide (URL below) it looks like there is a "Drive Information" option in that program that likely will query the drive with vendor specific ATA commands.

http://www.seagate.com/staticfiles/support/seatools/user%20guides/SeaTools_for_Windows.EN.pdf

Unfortunately when I boot Windows the ST31500341AS can only be accessed through an add-in PCI HBA card because my mobo has no native SATA ports. I'm currently not employed and will not be buying a new mobo in the next month or two. SeaTools for Windows cannot see the ST31500341AS when it is connected to that particular PCI HBA. With the appropriate drivers installed Windows sees the drive just fine.



The bottleneck is, of course, the PCI HBA card.



I connect my SATA boot drive to a PATA connector on the mobo via a SYBA SD-ADA50016 IDE <-> SATA bi-directional adapter that uses the Sunplus SPIF223 chipset. As far as I'm concerned it works just fine. Yes, it's a bit slow. Some folks on this board will likely laugh at it ... no big deal.

I unplugged the boot drive and plugged in the ST31500341AS to the SD-ADA50016. I then booted from a BartPE ISO CD and after that installed HD Tune Pro v4.60 which complained about the lack of a dll which I then copied into the same dir as HD Tune. After that it ran the cache test fine.

I hooked things up like this to (maybe) get a little more speed on the cache test and to check the results of the first test.

This time the speed of the transfers coming out of the cache were about 70 MB/sec. The prior test on the PCI card HBA had gotten around 60 MB/sec. The screen capture from the second test is below. The first test also had a 15-20% speed dip between 2-4 MB ... similar to this test. I realize I don't need to have the "4KB align" option selected, but it shouldn't hurt.



As far as I can see the main result of the test is that the speed remains relatively constant even though the data transfer size increases beyond 32 MB. That says to me that the drive has more than 32 MB of cache. I suspect that a situation exists similar to the "cache swap" in the article at the URL below. That HDD was supposed to have 8 MB of cache, but on tear down was observed to have 32 MB cache. I've also heard of situations where folks were supposed to get 8 MB cache and instead got 2 MB cache.

http://www.storagereview.com/western_digital_scorpio_blue_review_750gb_wd7500bpvt



The drive is brand spanking new and under warranty. Maybe in a few years I'll peek at the "chip side" of the controller PCB, but not now.

I thank you for your time and your suggestions fzabkar.

Cheers, mate!

I still think that your Sunplus bridge board, or the motherboard's IDE interface, is the bottleneck for your cache test.

If you examine HD Tune's read benchmark, I believe you will find that yours flatlines at 60MB/s or 70MB/s, whereas a normal curve should look like the following:

http://dyski.cdrinfo.pl/benchmark/hdtun ... 86-108.png

I haven't seen any HD Tune cache test results, but I expect that a normal curve might look like this:


BTW, your SPIF223 bridge chip appears to support Ultra ATA 100/133:
http://au.alibaba.com/product/379515489 ... e-HDD.html
http://i00.i.aliimg.com/img/pb/651/578/ ... 51_776.jpg

_________________
A backup a day keeps DR away.

WDC5000AAKX