Shortstroked WD5001AALS-00E3A0 Caviar Black 500GB

[Franc Zabkar]

A Western Digital user recently suggested that his 500GB
WD5001AALS-00E3A0 Caviar Black drive may have been fitted with two
500GB platters which had been shortstroked:

http://www.tomshardware.com/forum/f...rint=0&numreponse=0&quote_only=0&new=0&nojs=0

He reports that his "benchmark read/write from HDtune shows maximum
read/writes as 129MB/sec and minimum as 97MB".

http://i54.tinypic.com/1zlr9f5.png

AISI, the subject drive's maximum transfer rate of ~130 MB/s is
consistent with a 500GB platter density. The number of zones also
appears halved, suggesting that the drive has indeed been
shortstroked.

Other WD5001AALS models (eg WD5001AALS-00L3B2) appear to have 2
platters with densities of 250GB per platter. My understanding is that
the data transfer rate increases with the square root of the platter
density.

Hence ...

data rate (00L3B2) / data rate (00E3A0) = sqrt (250GB / 500GB)

ie data rate (00L3B2) = 130 x sqrt (0.5) = 92 MB/s

This is consistent with the benchmark results.

I also believe that the subject drive may have only 3 active heads,
not 4. I say this because its performance graph shows 9 zones. The
performance graphs for drives with a full complement of 18 or 19 zones
indicate that the 9th zone ends at about the 64% capacity point. If
the subject drive has a capacity of 250GB per head, then ...

3 x 250 x .64 = 480GB

.... which would tend to confirm that only 3 heads are in use.

The only question I have is, if the drive is shortstroked, why isn't
its seek time significantly lower?

Here are several HD Tune read benchmark graphs:

WD5001AALS-00E3A0 - 129MB/s max, 97 MB/s min, 11.0ms access, 8 or 9
zones:
http://i54.tinypic.com/1zlr9f5.png

WD5001AALS-00E3A0 - 136MB/s max, 99 MB/s min, 12.1ms access, 9 zones:
http://s001.radikal.ru/i193/1007/b2/b9e907d1985f.jpg

WD5001AALS-00L3B2 - 92MB/s max, 44 MB/s min, 12.2ms access, 19 zones:
http://i77.photobucket.com/albums/j63/euro-kid/HDTune_Benchmark_WDC_WD5001AALS-00L.png

WD5001AALS-00L3B2 - 92MB/s max, 44 MB/s min, 12.3ms access, 19 zones:
http://www.overclock.net/attachment...ts-hdtune_benchmark_wdc_wd5001aals-00l3b2.png

WD5001AALS-00L3B - 97MB/s max, 44 MB/s min, 12.5ms access:
http://dyski.cdrinfo.pl/benchmark/hdtune/hdtune-832-85758-154.png

WD5001AALS-00L3B - 92MB/s max, 45 MB/s min (est), 12.6ms access, 18 or
19 zones:
http://www.uppicweb.com/x/i/iy/90dhd.jpg


- Franc Zabkar

 

[Arno]

A Western Digital user recently suggested that his 500GB
WD5001AALS-00E3A0 Caviar Black drive may have been fitted with two
500GB platters which had been shortstroked:

He reports that his "benchmark read/write from HDtune shows maximum
read/writes as 129MB/sec and minimum as 97MB".

AISI, the subject drive's maximum transfer rate of ~130 MB/s is
consistent with a 500GB platter density. The number of zones also
appears halved, suggesting that the drive has indeed been
shortstroked.

Other WD5001AALS models (eg WD5001AALS-00L3B2) appear to have 2
platters with densities of 250GB per platter. My understanding is that
the data transfer rate increases with the square root of the platter
density.

Hence ...

data rate (00L3B2) / data rate (00E3A0) = sqrt (250GB / 500GB)

ie data rate (00L3B2) = 130 x sqrt (0.5) = 92 MB/s

This is consistent with the benchmark results.

I also believe that the subject drive may have only 3 active heads,
not 4. I say this because its performance graph shows 9 zones. The
performance graphs for drives with a full complement of 18 or 19 zones
indicate that the 9th zone ends at about the 64% capacity point. If
the subject drive has a capacity of 250GB per head, then ...

3 x 250 x .64 = 480GB

... which would tend to confirm that only 3 heads are in use.

The only question I have is, if the drive is shortstroked, why isn't
its seek time significantly lower?

Seek time is only weakly dependent on the seek distance. Sure,
track-to-track is faster than full stroke, but half-stroke is not
that much faster than full-stroke.

Arno

 

[Franc Zabkar]

Seek time is only weakly dependent on the seek distance. Sure,
track-to-track is faster than full stroke, but half-stroke is not
that much faster than full-stroke.

Here is the benchmark graph for a "full stroke" drive:
http://i77.photobucket.com/albums/j63/euro-kid/HDTune_Benchmark_WDC_WD5001AALS-00L.png

If you extrapolate the bottom of the access time graph, the
full-stroke time is about 16 msec.

The half-stroke point would be where the transfer rate is half way
between the min (44MB/s) and max (92MB/s) values, ie somewhere between
the 65% and 70% mark. The access time at this point is between 12 and
12.5 msec. This is near enough to the same figure that we measure for
the "full-stroke" of the short-stroked drive.

- Franc Zabkar