Panasonic 512 MB and PNY
(Toshiba) 256 MB SD Cards
Until I bought my Palm Tungsten|T3, I lived
in Sony Memory Stick land for external media needs. The performance
amongst Memory Stick offerings holds very uniform. That's far
from true for Secure Digital (SD) cards. For an explanation
of the various memory card technologies, see here.
For their relative specifications, see here.
In my experience to date, combined with forum threads on the
net, no two SD card brands show greater performance disparity
than Panasonic and Toshiba card offerings.
Toshiba seems to enjoy an expanding presence
in the repackaging market. Many rebranded cards like the PNY,
DaneElec, Lexar, etc., are based on Toshiba underpinnings.
DaneElec and Lexar in particular used to be Panasonic strongholds.
Economics undoubtedly drive this move, as Toshiba-based cards
sell for less than Panasonic-based cards. But what do we give
up to save a few bucks? That's the real question addressed
in this article, and the answer might surprise you.
First, a word about testing SD card performance.
Results depend on many things independent of the card circuitry
itself. The card-testing program of choice these days is VFSMark.
It turns out that VFSMark results depend to some extent on
the free space on the card. For that reason, I tested the cards
when newly formatted using CardInfo. Different devices will
test differently due to OS and interface overhead. Cards of
different manufactured capacities will generally vary in performance
depending on the brand.
Also, VFSMark emulates standard Palm OS file
operations in that it uses small block memory transfers during
testing. This accurately predicts the performance of standard
Palm OS file functions, as well as HotSync performance to the
card. However, programs like CardExport use
large block transfers and can offer radically different performance.
In addition, VFSMark results are not absolute values, but are
relative to a Palm m500 with a Palm 16MB SD card. The final
score consists of an unweighted average, therefore one great
result in any area may skew the final score. As a result, total
VFSMark scores may not be representative of the user's experience
when comparing SD cards. You should always look at the individual
test numbers that go into the final score. I'll prove this
in a few paragraphs.
I say all this so that users can properly
interpret results they find on the net and how those results
might apply to their situation. The most applicable results
will be those that match your device, card manufacturer, and
card capacity. Match as many parameters as you can before taking
the information as applicable to your situation. Compendiums
of VFSMark test results for various cards and devices can be
and T3 results on a forum thread here.
For consistency with other tests you might
find on the net, I used VFSMark 1.1 to evaluate both cards
on a Palm Tungsten|T3 (400MHz CPU). I tested PNY
(Toshiba) 256 MB (ID SD256 under CardInfo) and Panasonic
512 MB (ID SH512 under CardInfo) SD cards. I ran the tests
a number of times on both cards to ensure consistency. I also
timed the test to add another point of comparison and to see
how close the overall score matches the user's experience.
243.1 MB when formatted
File Create: 42%
File Delete: 23%
File Write: 5%
File Read: 748%
File Seek: 1475%
DB Export: 24%
DB Import: 1074%
Record Access: 724%
Resource Access: 698%
This test took about 10 minutes and 52 seconds
MB when formatted
File Create: 555%
File Delete: 331%
File Write: 158%
File Read: 725%
File Seek: 1072%
DB Export: 302%
DB Import: 786%
Record Access: 724%
Resource Access: 662%
This test took about 38 seconds
From the summary VFSMark scores, it would
seem as if the user experience from both cards would be comparable.
Nothing could be further from the truth. I have enough use
on both cards to make some subjective assessments. The Toshiba
indeed reads screamingly fast, although in practical use the
Panasonic feels just as fast. However, note the disparity in
VFSMark test times. The Panasonic finishes the test over 10x
faster! My experience shows that the time to complete the test
more accurately represents the overall user experience with
the Toshiba card.
The first three tests (which involve writing)
on the Toshiba take almost all the test time, with the File
Write test taking the overwhelming amount of time in that segment.
Watching the File Write test progress provides an experience
akin to watching grass grow in the arctic. In practical use,
USB HotSyncing large files (~10 MB and larger) to the Toshiba
SD can literally take over an hour. As long as you're just
reading the card, e.g., playing MP3s or looking up data, the
card performs very quickly. When writing to it, the card is
excruciatingly slow. My old 66 MHz T665C wrote way faster to
its 128 MB Memory Stick (VFSMark 98) than my 400 MHz T3 does
to the Toshiba SD cards, and that just isn't right.
The exception comes through CardExport 1.14.
CardExport apparently writes in large blocks, and the Toshiba
chipset is apparently optimized for large block writes. To
compensate for the slow HotSyncing to the Toshiba card, I used
to copy my daily news in HandStory to the card using CardExport.
HotSync took about 5-10 minutes depending on the amount of
news, while CardExport took about 15 seconds or so.
On the other hand, the Panasonic writes the
fastest of all the cards I've seen tested on the net. Writing
to any medium is the most time consuming thing a computing
device does, which is why the write tests dominate the VFSMark
test time. The Panasonic's 38 seconds test time truly represents
the overall user experience. Writing to the card is very fast,
and HotSyncing large files proved reasonable (6.5 MB in 4 minutes)
even for that inefficient process. BackupBuddyVFS 2.15 took
as long for a complete, from-scratch backup (527 files) on
the Panasonic as a differential backup (about a dozen or so
files) on the Toshiba. Life is good again at Tanker Bob's.
When you buy third-party cards like PNY,
Lexar, SimpleTech, DaneElec, etc., the label doesn't tell the
whole story. Sometimes the only way to tell who made the card
requires putting it in a PDA and running something like CardInfo
on it. When I asked one vendor about a particular card's manufacturer,
they told me that the differences weren't significant! I politely
disagreed with them. If you want a particular chipset, ask
the vendor and ensure their return policy will cover you if
what you receive isn't what you expected. Been there, done
Several manufactures are
marketing new cards touted as enhanced in some way. They
go by various names—32x,
High Speed, Pro, Ultra, et al. The Kingston Elite Pro High
Speed SD cards reviewed here on
PDABuyersGuide furnishes just one example. Given the dominance
of write performance on the user's daily experience quality,
considerable weight should be given to VFSMark's File Write
test parameter. The overall VFSMark score doesn't necessarily
provide a realistic measure of user experience. That's not
to say that some of these cards aren't faster in some areas,
but they may not uniformly provide a better overall user experience.
Panasonic clearly dominates
high-performance user experiences on the Palm T3, and I suspect
other devices, with their standard SD cards. Write performance
tops every other SD card currently on the market by a wide
margin, while read performance remains very competitive.
You'll pay a bit more for a Panasonic than other “standard” cards,
but I found the performance improvement well worth the price
difference. Caveat Emptor!