Hi Friends,

I just purchased an 8GB MicroSD Class 6 (Transcend).

I browsed through the Internet to find out if there's a way to check/verify that I really purchased a Class 6 device. I found some tools that can do such:

Linux: F3 version 1.1.3 (a.k.a. Fight Flash Fraud) - used this in Ubuntu 10.10
Microsoft: h2testw version 1.4 - used this in Windows 7 32-bit Ultimate

I was unsatisfied with what I saw and thought that I purchased a "fake" MicroSD. So I tried to do an actual data transfer by copying a huge file over to the N900.

Here's a screenshot. At the left is the benchmarking tool; at the right is a screenshot of actual data transfer (tried to copy a 2GB-ish movie file. The N900 is attached to the computer in Mass Storage Mode, copied the file directly to the MicroSD).




I also checked the Transcend website to verify the details of the card. I used the Taiwan site (left side) and the Philippines site (right side). Both results seem consistent.



I'm confused and maybe someone can give me advice.

1. Which screenshot reflects the actual/correct transfer rate so I can say whether I got a genuine card or not?
2. Is the benchmarking tool "inaccurate"? I tried the Linux counterpart of the tool and got the same results (sorry I wasn't able to take the screenshot).

Your advice matters. It will help me determine whether I should return the MicroSD.Thanks!

EDIT: When I tried the benchmark test at first attempt, I did it by using the MicroSD adapter (so it would be the size of a regular SD Card) which is included in the product I purchased, then plugged it into my computer's card reader. I'm currently booted into Linux (Ubuntu 10.10) and this time, I'm doing the benchmark from where the MicroSD is directly plugged into the N900 instead. Currently in progress...I observe the transfer rate to be faster, so I think it could be either the card reader or the adapter that slowed it down. I'll let this finish and will update this post.

UPDATE 01: I just finished running the benchmark test for "write" speed. With this test, the MicroSD is inserted into the N900, and then the N900 is plugged into my desktop computer in Mass Storage Mode. Results were favorable.



Currently running the "read" speed benchmark. If the results will still be good, I may have to mark this thread as solved (keeping fingers crossed).

UPDATE 02: Here are the results. 4MB/s "read" speed. Hm...? That means my questions will have to stay as they are (unfortunately).



In Wikipedia, it says here though that a Class 6 would give about 6MBytes/s read/write: http://en.wikipedia.org/wiki/Secure_...#.C3.97_rating. I guess I have to run some more tests, for now I have to sleep. Perhaps someone can interpret this information better? Thanks in advance.

UPDATE 03: I just did another test today in Linux (Ubuntu 10.10) with the MicroSD in the N900, and then plugged the N900 to my desktop computer in Mass Storage Mode. Results are still the same as with Update 02 (9Mb/s read, 4Mb/s write). So I tried with Windows again (MicroSD in N900 then plug N900 to computer). Here are the results. I got 8.61Mb/a read, 9.09Mb/s write.



CONCLUSION/OBSERVATIONS:

1. Was able to verify the authenticity of the Transcend MicroSD card (Class 6, 8GB) I purchased by looking it up at the product's official website, as well as some speed benchmarking tools available for download.
2. Reported benchmarks/speeds of MicroSD seemed to differ based on application used, Operating System used, as well as how the MicroSD is plugged in for use (e.g. via adapter+card reader versus via direct attachment/plug to a device).
3. Summary of my test results:

Microsoft Windows

1. Plugged in to MicroSD-to-SD adapter+plugged in to Card Reader: 3.80 Mb/s Write; 4.42 Mb/s Read using H2testw application
2. Actual data copy/transfer (using a 2.48 GB video file): 9.70 Mb/s
3. MicroSD inserted to a device (N900), and then device plugged directly to computer (USB port at back of computer) via Mass Storage Mode:

Linux (Ubuntu 10.10)

MicroSD inserted to a device (N900), and then device plugged directly to computer (USB port at back of computer) via Mass Storage Mode: 9.12Mb/s write; 4.12Mb/s read

At this point I can finally mark this thread as "Solved (or closed)". The important thing is that I was able to share my experience and hope that this may help others as well.

Friends,

Care to share your results? You may want to use the tools attached.

First Attachment: F3 version 1.1.3 (Fight Flash Fraud) for Linux users. Please follow the instructions here: http://oss.digirati.com.br/f3/. I'm posting the instructions for compiling just to be sure:

To use the application (this is all command-line only, no GUI), see below. Make sure that you are in the directory where you "unzipped" or installed F3!

For Write Test:

For Read Test:

Second Attachment: h2test2 version 1.4 for Microsoft users. Just unzip the file to your directory of choice and run the program; no installation needed.

NOTES:

1.(I had to add this based on epithaph's comment): benchmarking tools will fill up your storage device when doing a "write" test; it is intentional because these files will be needed when you run the "read" test. You can delete those files later after you finish your benchmarking.

2. when we say "name of storage device", it means you have type the label of the storage device. For example, you labeled your MicroSD as "boomboom"...then when you run the command for F3 (the Linux-based benchmarking program), this is how it should be typed:

For the write test

For the read test

For the Microsoft-based program, no need to worry about commands since it's GUI-based.

Thank you for sharing but f3read isn't working it gives me:

./f3read / or ./fread /media/mmc1

Data OK: 0.00 Byte (0 sectors)
Data LOST: 0.00 Byte (0 sectors)
Corrupted: 0.00 Byte (0 sectors)
Slightly changed: 0.00 Byte (0 sectors)
Overwritten: 0.00 Byte (0 sectors)
Reading speed: 0.00 Byte/s

I've followed your instruction. Any idea?

Hi Friend,

I got that too the first time I tried. Apparently we have to use the command for "writing" data on the card first (./f3write) so that there will be data to be "read" once you try the command for "reading" (./f3read). Also just a quick reminder/disclaimer from the website where I got F3:

I did the tests on a newly-purchased MicroSD card...I only had one file in the card which is the 2GB video file I used in the testing as mentioned in Post #1, although when I re-ran the test without deleting the video file, it didn't seem to get overwritten. Anyway the author of the file said that the program will overwrite/delete files only if the card has a problem (marked with underscore above). Hope this helps.

EDIT: did you happen to "rename" or "put a label" on the MicroSD? I labeled mine as "MICROSD" as mentioned in the command I used as an example.

Right now I'm testing it on my /home/user folder:
./f3write .
Free space: 1.48 GB
Creating file 0001.fff ... OK!
Creating file 0002.fff ...

it took very long time. I will report back the results. So far my writing speed:
./f3write .
Free space: 1.48 GB
Creating file 0001.fff ... OK!
Creating file 0002.fff ... OK!
Free space: 0.00 Byte
Writing speed: 1.96 MB/s

My reading speed:
./f3read .
SECTORS ok/corrupted/changed/overwritten
Validating file 0002.fff ... 1004304/0/0/0
Validating file 0001.fff ... 2097152/0/0/0

Data OK: 1.48 GB (3101456 sectors)
Data LOST: 0.00 Byte (0 sectors)
Corrupted: 0.00 Byte (0 sectors)
Slightly changed: 0.00 Byte (0 sectors)
Overwritten: 0.00 Byte (0 sectors)
Reading speed: 13.06 MB/s

The tool is eating all my free space in /home/user:
df -h .
Dateisystem Size Used Avail Use% Eingehängt auf
/dev/mmcblk0p2 2,0G 2,0G 0 100% /home

Without my super-duper tweaks my read speed is much more faster:
./f3read .
SECTORS ok/corrupted/changed/overwritten
Validating file 0002.fff ... 1004304/0/0/0
Validating file 0001.fff ... 2097152/0/0/0

Data OK: 1.48 GB (3101456 sectors)
Data LOST: 0.00 Byte (0 sectors)
Corrupted: 0.00 Byte (0 sectors)
Slightly changed: 0.00 Byte (0 sectors)
Overwritten: 0.00 Byte (0 sectors)
Reading speed: 19.42 MB/s

And again with my super-duper tweaks:
./f3read .
SECTORS ok/corrupted/changed/overwritten
Validating file 0002.fff ... 1004304/0/0/0
Validating file 0001.fff ... 2097152/0/0/0

Data OK: 1.48 GB (3101456 sectors)
Data LOST: 0.00 Byte (0 sectors)
Corrupted: 0.00 Byte (0 sectors)
Slightly changed: 0.00 Byte (0 sectors)
Overwritten: 0.00 Byte (0 sectors)
Reading speed: 18.70 MB/s

Nice, you're on the right track.

Yeah it will indeed take a long time (especially if you have a MicroSD card larger than mine, e.g. 16GB or so). The way I understand it, the program tries to "write" data first to the card and does a "write" benchmark. And then you'd run a "read" benchmark where the program tries to "read" the data that was initially written. And then it will post the results to you.

Based on my experience in Post #1, running the benchmarking test in a Microsoft-based OS gave me higher numbers compared to F3 (Linux-based) when it comes to "read" speed. No matter what I did, F3 always gave me an average of 4 Mb/s "read" speed (although "write" speed seemed okay).

EDIT: I didn't notice you ran the command on /home/user...you are benchmarking the speed of the eMMC? (32GB, particularly the "Home" partition)?

EDIT 02: Hi Friend, saw your edit (posting the results). You may want to try using another tool, like the Windows-based tool so you can compare just like what I did. I got slower benchmarking results in the Linux-based program compared to Microsoft-based program.

EDIT 03: Your read speed is fast.

EDIT 04: Yes, it will indeed fill up the storage that you tried the benchmarking on. That's the intention of these tools as mentioned by the author of the program:

1. "Write" will place files to fill up the storage device and at the same time, will measure how fast it can "write" files to the storage.
2. "Read" will read the files that were initially "written" and will measure how fast it can "read" those files.

You can safely delete the new files that were "written" or "created" by the program after you finished your benchmarking, by the way. It's easy to identify those files that you need to delete; they are all in 1GB increments. The files are named:

...and so on.

Hi Epitaph,

Thanks for sharing your results. Will try the same on my eMMC too when I have the time; so far I only tested my MicroSD as I was driven by anxiety on whether the card I purchased was fake or not.

I have deleted the file 0002.fff because my disk is full. The benchmark is working on 0001.fff too. It's not need to make the disk full with this chunks. Why did he makes it anyway? Or is there an option?

I think there needs to be at least one file that the program has created so it's okay to delete the other created files. Not sure if there is an option to change that. Here's what the author said, from his website:

So I guess it's required to have at least one file that was created by "f3write".

Hi, I'm not sure what your are talking about but I just follow your suggestion. I didn't have time to read the website. Isn't there not a simple program to benchmark? I want to test time dd if=/dev/null of=/dev/mmcblk0p1 count=512 bs=512 because I remember it is a good benchmark, too. PLEASE don't try this yourself. It's not tested!. What do you think is the difference? Me just to lazy to write my own shell-script?

EDIT:Here is a good tutorial how to use dd non-destructive and dma: http://wiki.linuxquestions.org/wiki/..._of_your_disks

EDIT: dd iflag=direct if=/dev/mmcblk0p1 of=/dev/null bs=512 count=100000
100000+0 Datensätze ein
100000+0 Datensätze aus
51200000 Bytes (51 MB) kopiert, 41,2531 s, 1,2 MB/s

dd if=/dev/mmcblk0p1 of=/dev/null bs=512 count=100000
100000+0 Datensätze ein
100000+0 Datensätze aus
51200000 Bytes (51 MB) kopiert, 3,17645 s, 16,1 MB/s

dd if=/dev/mmcblk0p1 of=/dev/null bs=512 count=900000
900000+0 Datensätze ein
900000+0 Datensätze aus
460800000 Bytes (461 MB) kopiert, 24,7127 s, 18,6 MB/s

dd if=/dev/mmcblk1p1 of=/dev/null bs=512 count=900000
900000+0 Datensätze ein
900000+0 Datensätze aus
460800000 Bytes (461 MB) kopiert, 34,2288 s, 13,5 MB/s

Blocksize

dd if=/dev/mmcblk0p1 of=/dev/null count=1
1+0 Datensätze ein
1+0 Datensätze aus
512 Bytes (512 B) kopiert, 0,0335083 s, 15,3 kB/s

dd if=/dev/mmcblk1p1 of=/dev/null count=1
1+0 Datensätze ein
1+0 Datensätze aus
512 Bytes (512 B) kopiert, 0,00357056 s, 143 kB/s

Unfortunatley dd works only when you have bash 4.2 installed!

Here another benchmark without my tune-up tools:

/dev/mmcblk0p1
900000+0 Datensätze ein
900000+0 Datensätze aus
460800000 Bytes (461 MB) kopiert, 25,3318 s, 18,2 MB/s

/dev/mmcblk1p1
900000+0 Datensätze ein
900000+0 Datensätze aus
460800000 Bytes (461 MB) kopiert, 28,351 s, 16,3 MB/s

here with my tune-up tools:

900000+0 Datensätze ein
900000+0 Datensätze aus
460800000 Bytes (461 MB) kopiert, 21,748 s, 21,2 MB/s
900000+0 Datensätze ein
900000+0 Datensätze aus
460800000 Bytes (461 MB) kopiert, 27,0826 s, 17,0 MB/s