despite what the guy replying is saying, sadly rumour has it on storagereview forums, that you will get 160gb (slower) platters, NOT 188's stroked back

So the latest 7200.10model of the 400gb is still sadly 3x133 platters just like the 500gb is 4x125gb platters :(

I wonder what it can do to a person if 750GB of data was lost...heheh.

And based on the graphs, it bests the raptor in lots of cases, which means that this 7200rpm drive would do way better than a 15k drive based on previous tests on the raptor/scsi drives. something is not right.

This is an interesting article about a potentially very interesting drive, but it's stunning that you did all of this work benchmarking hard drives for various application performance scenarios without running any basic I/O tests: sequential and random reads and writes for large and small block sizes. Your testing would be great if I were running the applications you benchmarked, but without any of the basic metrics, it's impossible for me to determine whether this drive would fit my needs.

I guess I'll have to buy my own drive and run my own tests. :-)

Those iPeak graphs look rather odd. Some disks inexplicably shoot out in the lead on one test but not any of the others, yet for many tests it's a dead heat across the board. None of the disks consistently beats the pack, either, yet these freak standouts?

What's up with that?

Not that it's important or anything ('cause it's not, really), but for the people who are curious: the arrows in the drawing do not represent a 1 or a 0, but just magnetic flux reversals. It takes many flux reversals in a particular pattern to define a series of bits. I'll get this wrong, but I'll say the minimum recognizable clump of data is four bits, and so there are 16 flux reversal patterns defined that have to be recognized on the fly.

The flux reversals fly by the heads so fast and create such an indistict electrical profile that to map them one-one with data bits would be impossible. The signal coming off the platter looks like FM noise, and it takes a very fast DSP (on the drive) to do the translation from noise to data bits.

Someone with real knowledge of current PRML can correct me here. I just wanted to point out that the up arrows and down arrows are not ones and zeros, and the ratio of flux reversals to bits is greater than one. It averages something like 5:1, IIRC.

I have always noticed that in hard drive reviews, game load time tests seem to be rather inconsistant. Has anyone ever thought of why this is. I guess some games may load different sized files (a lot of small files or a few large). Maybe it has to do with where the file is located physically on the hard drive. I don't know if that would have a measurable effect or not.

Just a thought.

Is there any particular reason y'all don't show the IOMeter response time graphs with a log y-axis? It seems like it would help show off the differences between the drives, which are kinda obscured by the exponential scaling...

Makes me wonder if I should cancel my order on my 150gb raptor O.o

I'm confused: was NCQ support enabled or not (for those drives that support it)? We know that it can make a significant difference (good and bad), and it's not clear to me which tests (if any -- IPEAK?) had it enabled. I'd actually like to see all the drives tested with NCQ both on and off but that would double the testing load....

Diamondmax 10 is highlighted on the Adobe Photoshop graph.

Does anyone know if 320GB 7200.10 uses a 188GB, or a 160GB platter?

160GB would make more sense (ie. 160 x 2 = 320), but someone told me it's using a 188GB platter...?

Can anyone verify this?

Thanks ;)

To NCQ? Or not to NCQ?

Unfortunately, the review doesn't give an answer to that. :(