First, I have the map complete with data overwritten in order to be able to test in "the steady" state.
- Speed Test:
Has been tested with the Linux tool "dd" and block size 1M, with direct access to the device (/ dev / sd * 1), the command "sync" has been executed partly shortly after start of writing to no cache to write synchronously, for Compared were a few tests were done once with a synchronous gemountetem exFAT reached at the similar rate.
In my Transcend RDF5 USB3.0 Card Reader (Card Reader Maximum: 90MB / s read, 60MB / s write) I could reach following sustained transfer rates:
Writing 16MB with sync: 37 MB / s, 14.9 MB / s,
Reading 128MB: 81.1 MB / s
3x writing 128MB without sync: 24.1 MB / s, 32.9 MB / s, 25.3 MB / s
3x letter from 128MB to sync: 9.1 MB / s, 12.9 MB / s, 17.5 MB / s
Reading 2GB: 83.9 MB / s
3x 2GB writing without sync: 37.1 MB / s, 29.6 MB / s, 29.0 MB / s
3x letter from 2GB to sync: 19.3 MB / s, 15.3 MB / s, 18.7 MB / s
unaligned (unaligned) letter (ext2) of 16MB with sync: 700 kB / s, 1.1 MB / s, 980 kB / s
Conclusion: Under ideal conditions therefore seem 37MB / s write rate to be reached, but surely that's never. Reading, however, is great!
- Coarse reliability test
Since I schonmal Fake cards had been able to maintain, unfortunately, in real only 2GB data but could be formatted as 16GB, I have also tested:
I have the card completely filled with copies of Debian Linux DVD ISOs and other several small writes and eject / reinsert again all checksums tested, all of which were 100% correct. Long-term tests are still pending, I hope the card maintains a while ...