MaximumPC 2008 11

(Dariusz) #1

44 |MAMAMAXIMXIMXIMXIMUUUUMMPPPCC|NOV 08 |www.maximumpc.com


RiData’s 64GB SSD uses an MLC design to
pack more data onto its fl ash memory chips.
We like how that makes the drive cheaper
than the majority of SSDs on the market.
What we don’t like is how the Ultra-S Plus
illustrates the performance losses wrought by
using this technology instead of a speedier
SLC design.
The Ultra-S Plus was able to overtake the
fastest hard drive we’ve tested—Western
Digital’s Velociraptor—in two of our bench-
marks: a random access read measurement
and the overall PCMark Vantage score.
Neither win came as a surprise. Because
hard disk drives suffer lag while the drive
arm moves to the proper location on the
disk, flash memory consistently outperforms

magnetic storage in random access read
speeds. This helped in PCMark Vantage
because the app’s eight individual
benchmark traces favor read
performance and random
access reads.
The device’s
horrible write
performance—
including an aver-
age random access
write speed of 248 painful
milliseconds—was enough to
drag its PCMark Vantage score
below that of all the other SSDs in
this feature. And the Ultra-S Plus took
more than 1.5 times longer to complete
our real-world Premiere test than the fastest
SSD we tested, Memoright’s GT-Series 64GB.
RiData’s SSD operates over a SATA
3Gb/s interface, although our initial round
of interface speed benchmarks made this
drive appear to operate over a bridged PATA
connection. We believe that the drive’s MLC
flash chips threw off our speed tests at first.
But the fact that this SSD gave us such poor
read speeds over a SATA 3Gb/s interface

doesn’t paint a pretty picture for this device.
Indeed, you get what you pay for with RiDa-
ta’s SSD—we’d much rather have a $300 hard
disk drive instead of this solid state drive.

Super Talent’s 64GB SSD must be using the
exact same hardware as RiData’s Ultra-S
Plus 64GB. If not, then the similarities be-
tween these drives are an amazing coinci-
dence. We recorded identical random access
read times for both, an underwhelming .39
milliseconds. Both drives’ PCMark Vantage
scores were within one-third of one percent
of each other, and they varied by just two
seconds in our uncompressed AVI file-
creation test.
If these two MLC-based drives are in-
deed brothers in arms, then they’re the two
drunken soldiers stumbling around at the
rear of the SSD brigade. Like the RiData, the
Super Talent’s performance is unacceptable,
even given its low price. While the Super
Talent drive overtakes our Western Digital
Velociraptor in the real-world PCMark Van-

tage test, we’d be terrified to use this drive
as the primary storage for our operating
system. Its random access read scores are
swift, but this drive’s random access write
performance is atrocious: It was more than
7,000 percent slower than a Velociraptor in
our tests!
This drive would rock if we only needed
to read information from it, but the SSD’s
write speeds are simply too slow. There’s an
inexplicably large gap between the Super

Talent’s slowest and fastest sustained read
speeds: 14.7MB/s, as recorded by h2benchw.
This doesn’t make much of a diff erence in real-
world performance, but it’s certainly greater
than the 2MB/s to 3MB/s diff erence, at most,
that we’re used to seeing from SSDs.
Like the Ultra-S Plus, the Super Talent
DX’s sustained read speeds outpace the
theoretical interface speed measurement.
But this is clearly indicative of a benchmark
snafu, as the Super Talent was unable to fill
the pipe of its SATA 3Gb/s connection in our
real-world tests.

This 2.5-inch drive spit out
the slowest fi le-write times
of all the SSDs we tested.

The Masterdrive DX’s
write speeds suffer due
to its MLC fl ash chips.

VERDICT

$300 MSRP, http://www.ritekusa.com^5


RIDATA ULTRA-S PLUS 64GB

VERDICT

$400 MSRP, http://www.supertalent.com^4


SUPER TALENT MASTERDRIVE DX

Super Talent


Masterdrive DX


This drive’s super talent


is slowness


RiData Ultra-S


Plus 64GB


Lower cost equals lesser


performance


flashflood

Free download pdf