I can tell you this much: PNY is the worst card I have ever bought. It's the only one I can put in my camera for a guaranteed slow-down. Also, if I put a movie on it and try to play back, it skips frames like mad.
I'm not really sure how realistic PM is at benchmarking.
If anyone is considering buying a PNY card, take my advice and walk away!
__________________
"I'm a perfectionist, but not very good at it yet."
Alright already. Some new test results in a card reader - below are the times needed to copy 919MB (964108288 bytes) worth of MP3 files i.e. 6.3 MB average file size from PC to card. All cards formatted without backup FAT.
Since posting the first spreadsheet I have retested almost all of my cards with Spb Benchmark 1.6. (The ratings in that first spreadsheet were obtained with Spb B 1.1.) And I have been experimenting with formatting options and added three new cards: CF Transcend 8GB, CF Extreme III 4GB, SD Transcend 4GB.
The new spreadsheet is attached here and in my updated first post to this thread.
In the Axim, according to Spb Benchmark's Speed Index, the RiData 2GB 150x and Transcend 4GB 150x are fastest closely followed by the SanDisk Extreme III and Ultra II. The SDs are faster than the CFs. At the low end the MicroDrive performs very well compared to the other CF cards. I cannot say that I have felt real differences when using these cards in the PDA because I have not devised a systematic test for that.
Anyone reading this who has an idea how to test real world card speeds in a PPC in a reproducable way, speak up.
In a card reader, according to HDBench as well as real life tests, the Extreme III cards - both SD and CF - rule, followed by the RiData. The Transcend 150x is a distant third, not much better than the 70x and 60x cards. Forget about the MicroDrive.
Of the cards I have tested the RiData has good or best read and write speeds both in the Axe and a card reader, and costs less than the Extreme IIIs.
I have been using various cards for games that play live music/video, and for a GPS program that does error checking on the map database during startup. I can verify that actual read usage is proportional to the Pocket Mechanic read benchmarks.
There is something strange with the I/O of different devices for different cards. In general, faster cards are faster in a Pocket PC, but actual performance can defy the speed rating on the card. I look forward to benchmarking my new A-Data 2GB 50X SD card and comparing it to the 150X of the same brand.
For cards you use for pure reading on your pocket PC, for instance multimedia, all you have to worry about is good read performance.
But for a card you install programs on, consider write speed as well. If you close a program and turn off the device too fast before the program saves all it's data, you could have corrupted information on the flash media. I always leave my PPC "cool down" at least 15 seconds after closing programs, since then I don't have corruption errors.
__________________
"I'm a perfectionist, but not very good at it yet."
I'm using OCN5 with seamless USA(48)+Canada (>1GB). It's from the time I click on the program icon until I have access to the program. I'm not sure other GPS programs have a large enough map file to compare.
__________________
"I'm a perfectionist, but not very good at it yet."
OCN5 maps seem to be very small. Mapopolis Germany alone is north of 300MB. Garmin TX alone is over 100MB. I think I can make Garmin North America maps 2GB and larger. It should give me something to test. However, while this does seem to be a practical test, it is still not the reproducible test which I have in mind.
How is the general performance? Does it generally feel faster and more peppy playing stuff like video? My SanDisk Ultra II lags sometimes and I'm looking for something that can read faster. If you can confirm these results in real application usage, my money will be spent. Otherwise, I'll just wait until the Extreme III's get a bit cheaper.