Page 8 of 13

Re: SD Card Benchmarks

Posted: Wed Dec 18, 2013 6:57 am
by lingon
I wonder if someone with the Raspberry Pi NOOBS SD-card by Samsung MMCTR08G3ACH-QNJMK has run CrystalDiskMark or IOzone benchmarks on it? The announcement of the NOOBS card at
http://www.raspberrypi.org/archives/4536
says the card should be fast for random I/O:
Fast, pre-programmed, high-capacity cards like the Samsung ones we’re bundling (and which have turned out to be our favourite cards in testing; they’re optimised for random read/write behaviour, unlike many cards which are designed for the large continuous reads and writes that digital cameras make) have been the best-selling Pi accessory offered by our partners since launch, so we’re expecting a lot of you to take us up on this offer.
There are probably quite a few of these cards available as you can buy them separately, but I could not find any benchmark results for exactly the model name MMCTR08G3ACH-QNJMK. It would be interesting to know how this card performs compared to the fastest SanDisk models?

Re: SD Card Benchmarks

Posted: Fri Dec 27, 2013 9:34 am
by lingon
I found this link with CrystalDiskMark results for a SD-card with a similar but not exactly same type at
http://r-2ch.com/t/smartphone/1369656716/
The NOOBS card is MMCTR08G3ACH-QNJMK and the benchmarked card is MMCTR08G3ACH-MM.
The result for the MMCTR08G3ACH-MM card is not so impressive for random writes:
-----------------------------------------------------------------------
CrystalDiskMark 3.0.2 Shizuku Edition x64 (C) 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]
Sequential Read : 66.345 MB/s
Sequential Write : 6.409 MB/s
Random Read 512KB : 62.654 MB/s
Random Write 512KB : 5.640 MB/s
Random Read 4KB (QD=1) : 5.585 MB/s [ 1363.5 IOPS]
Random Write 4KB (QD=1) : 0.095 MB/s [ 23.3 IOPS]
Random Read 4KB (QD=32) : 5.199 MB/s [ 1269.2 IOPS]
Random Write 4KB (QD=32) : 0.090 MB/s [ 21.9 IOPS]
Test : 1000 MB [H: 0.0% (0.0/7515.0 MB)] (x5)
Date : 2013/08/10 15:04:19
OS : Windows 7 Professional SP1 [6.1 Build 7601] (x64)
If someone has the Raspberry Pi NOOBS card MMCTR08G3ACH-QNJMK, then it would be interesting to see CrystalDiskMark or Iozone benchmark results posted here.

Re: SD Card Benchmarks

Posted: Tue Jan 14, 2014 10:15 am
by mikerr
8GB Raspberry PI NOOBS card:

Image
http://imagizer.imageshack.us/v2/800x600q90/9/8qhh.jpg

To save you squinting at the image, the card markings are:

MMCTR08GUBCH-ML
D HTIA02GB 343
Made in KOREA

Benchmarks looking really good, beating the sandisk !

Re: SD Card Benchmarks

Posted: Tue Jan 14, 2014 11:42 am
by jdb
The card goes a lot faster than that for block reads...

Note that there are two limiting factors to the top read speed of a suitably fast SD card:

- The original SD bus speed (HS mode) maxes out at about 21MB/s (minus a bit of overhead).
- The USB2.0 throughput on most hardware is limited to less than 480Mbps for a single device bulk endpoint (typical 25-30MB/s)

I have a USB3.0 card reader that supports all the UHS-I high throughput modes and the NOOBS card can do 45-ish MB/s easily.

I will spin up crystaldiskmark this evening and give it a go.

Re: SD Card Benchmarks

Posted: Tue Jan 14, 2014 9:09 pm
by lingon
mikerr wrote:8GB Raspberry PI NOOBS card:

MMCTR08GUBCH-ML
D HTIA02GB 343
Made in KOREA

Benchmarks looking really good, beating the sandisk !
Thanks for sharing this result! I agree that it looks like a very good card for random writes.

Re: SD Card Benchmarks

Posted: Tue Jan 14, 2014 9:32 pm
by Richard-TX
Toshiba Exceria (Red Label)

Re: SD Card Benchmarks

Posted: Tue Jan 14, 2014 9:44 pm
by Richard-TX
exceria green label (32gb)

Re: SD Card Benchmarks

Posted: Wed Jan 15, 2014 12:36 am
by jdb
NOOBS 8GB SD card w/ USB3.0 reader

MMCTR08GUBCH-ML

OK so sequential read isn't as big as I thought - that'll teach me to trust win7 copy dialogue boxes
Noobs8GB.png
Noobs8GB.png (43.54 KiB) Viewed 27945 times
Kingston Class 10 card
Kingson class 10.png
Kingson class 10.png (43.26 KiB) Viewed 27945 times
Wow. That's a lot slower than I expected.

Re: SD Card Benchmarks

Posted: Wed Jan 15, 2014 12:52 am
by blachanc
welcome to the club... (re your kingston class 10)....


Buy an 8GB NOOBS SD card for £5

in the comment section: ;)
liz on
January 10, 2014 at 11:08 am said:
Yes. (Anyone who buys a Class 10 card for their Pi just because the number 10 is higher than the number 4, deserves everything they get.)

Re: SD Card Benchmarks

Posted: Wed Jan 15, 2014 12:57 am
by milhouse
jdb wrote:Wow. That's a lot slower than I expected.
Try reducing the size of your test data set, eg. try 50MB instead of 100MB.

I found that the same flash memory (USB3 memory stick in this case) could return good results at 50MB, even 100MB, but performance fell off a cliff when testing with a data set larger than 500MB. Maybe your Kingston card is suffering at only 100MB, though I've no idea why this should be the case (maybe an anomaly of the testing software, or it really is a poorly performing card?)

Re: SD Card Benchmarks

Posted: Wed Jan 15, 2014 5:55 am
by lingon
jdb wrote:NOOBS 8GB SD card w/ USB3.0 reader

MMCTR08GUBCH-ML
Thanks for sharing the NOOBS card USB3.0 results as well!

Re: SD Card Benchmarks

Posted: Wed Jan 15, 2014 9:02 am
by jdb
blachanc wrote:welcome to the club... (re your kingston class 10)....


Buy an 8GB NOOBS SD card for £5
I know, I just wasn't expecting the class 10 to be *that* terrible.
milhouse wrote:
Try reducing the size of your test data set, eg. try 50MB instead of 100MB.

I found that the same flash memory (USB3 memory stick in this case) could return good results at 50MB, even 100MB, but performance fell off a cliff when testing with a data set larger than 500MB. Maybe your Kingston card is suffering at only 100MB, though I've no idea why this should be the case (maybe an anomaly of the testing software, or it really is a poorly performing card?)
I repeated the test with 50MB test set size - no difference.

Edit: I knew I had seen these cards do faster than that: My desktop PC only has 2x USB3.0 ports and I plugged the reader into the wrong one :lol:

It appears there are two types of NOOBS card in the wild. The first is the class 4 variant but the version on sale right now is the class 6 variant.
Noobs class 6 8GB.png
Noobs class 6 8GB.png (44.8 KiB) Viewed 27869 times
Older noobs 8GB.png
Older noobs 8GB.png (48.43 KiB) Viewed 27869 times

Re: SD Card Benchmarks

Posted: Thu Jan 16, 2014 9:06 am
by StevenP
Here are the results from my SanDisk Extreme 8GB Class 10, 45MB/s, (SDSDX-008G-X46), taken over a USB2 port on a Windows 7 PC:
-----------------------------------------------------------------------
CrystalDiskMark 3.0.3 x64 (C) 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 19.195 MB/s
Sequential Write : 19.135 MB/s
Random Read 512KB : 18.952 MB/s
Random Write 512KB : 14.892 MB/s
Random Read 4KB (QD=1) : 4.089 MB/s [ 998.4 IOPS]
Random Write 4KB (QD=1) : 1.590 MB/s [ 388.2 IOPS]
Random Read 4KB (QD=32) : 4.642 MB/s [ 1133.3 IOPS]
Random Write 4KB (QD=32) : 1.727 MB/s [ 421.6 IOPS]

Test : 50 MB [H: 0.0% (0.0/7569.5 MB)] (x3)
Date : 2014/01/16 7:15:01
OS : Windows 7 Enterprise SP1 (x64)
sandisc 8GB extreme crystalDiskMark results 16Jan2014.png
SanDisk Extreme 8GB Class 10 45MB/s ( SDSDX-008G-X46 )
sandisc 8GB extreme crystalDiskMark results 16Jan2014.png (49.51 KiB) Viewed 27820 times
I believe the USB2 port I used limits the maximum speeds I can measure to 19MB/s.

The sequential write and 512kB random write speeds are 2-3 times better than the NOOB class 6 card. The 4kB random read and write speeds are similar to the NOOBS.

So I think I prefer this card to the NOOBS class 6 because of the faster large file write speeds, everything else being similar. It certainly performs very well in XBMC usage. It cost me 11 Euro from amazon.de.

Cheers,
StevenP

Re: SD Card Benchmarks

Posted: Sat Jan 18, 2014 5:49 pm
by jdb
StevenP wrote:Here are the results from my SanDisk Extreme 8GB Class 10, 45MB/s, (SDSDX-008G-X46), taken over a USB2 port on a Windows 7 PC:
-----------------------------------------------------------------------
CrystalDiskMark 3.0.3 x64 (C) 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 19.195 MB/s
Sequential Write : 19.135 MB/s
Random Read 512KB : 18.952 MB/s
Random Write 512KB : 14.892 MB/s
Random Read 4KB (QD=1) : 4.089 MB/s [ 998.4 IOPS]
Random Write 4KB (QD=1) : 1.590 MB/s [ 388.2 IOPS]
Random Read 4KB (QD=32) : 4.642 MB/s [ 1133.3 IOPS]
Random Write 4KB (QD=32) : 1.727 MB/s [ 421.6 IOPS]

Test : 50 MB [H: 0.0% (0.0/7569.5 MB)] (x3)
Date : 2014/01/16 7:15:01
OS : Windows 7 Enterprise SP1 (x64)

I believe the USB2 port I used limits the maximum speeds I can measure to 19MB/s.

The sequential write and 512kB random write speeds are 2-3 times better than the NOOB class 6 card. The 4kB random read and write speeds are similar to the NOOBS.

So I think I prefer this card to the NOOBS class 6 because of the faster large file write speeds, everything else being similar. It certainly performs very well in XBMC usage. It cost me 11 Euro from amazon.de.

Cheers,
StevenP

Interesting: so there is a card out there that has at least some optimisation for random read/write.

But consider that the NOOBS SD card is £4 + shipping (less than 11€).

http://swag.raspberrypi.org/collections ... gb-sd-card

Note that the vast majority of workload for a typical desktop OS is random read with a few random writes - interpreting how "fast" a particular card is for an OS workload is a non-trivial estimation.

Re: SD Card Benchmarks

Posted: Fri Feb 07, 2014 10:46 pm
by olso4539
I just received on of the class 4 SDHC NOOBS cards along with my raspberry pi ordered from Element 14 (Newark). The performance I'm seeing is similar to that posted above.

Re: SD Card Benchmarks

Posted: Wed Feb 12, 2014 8:41 pm
by lingon
I measured one of the Raspberry PI NOOBS 8 GB uSD cards that I received today in a native SD cardreader in a HP 4330s laptop.
The cards are marked:
MMCTR08GUBCH-ML
K PSX773ZV 351
Made in PHILIPPINES
These are the CrystalDiskmark results for one of my cards:
-----------------------------------------------------------------------
CrystalDiskMark 3.0.1 x64 (C) 2007-2010 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 46.363 MB/s
Sequential Write : 7.573 MB/s
Random Read 512KB : 44.966 MB/s
Random Write 512KB : 5.665 MB/s
Random Read 4KB (QD=1) : 6.647 MB/s [ 1622.9 IOPS]
Random Write 4KB (QD=1) : 1.832 MB/s [ 447.4 IOPS]
Random Read 4KB (QD=32) : 9.203 MB/s [ 2246.9 IOPS]
Random Write 4KB (QD=32) : 2.545 MB/s [ 621.4 IOPS]

Test : 50 MB [D: 17.5% (1314.2/7519.0 MB)] (x5)
Date : 2014/02/12 22:10:59
OS : Windows 7 SP1 [6.1 Build 7601] (x64)

Re: SD Card Benchmarks

Posted: Sat Mar 08, 2014 6:03 pm
by RazzpiNoob
This card is 10x-100x faster on small random writes than all my other SD cards and USB sticks. I chose it on a whim because the NOOBS card is a Samsung.

Samsung 16GB PLUS Micro SDHC
-----------------------------------------------------------------------
CrystalDiskMark 3.0.3 Shizuku Edition x64 (C) 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 19.695 MB/s
Sequential Write : 13.158 MB/s
Random Read 512KB : 18.955 MB/s
Random Write 512KB : 17.252 MB/s
Random Read 4KB (QD=1) : 4.465 MB/s [ 1090.1 IOPS]
Random Write 4KB (QD=1) : 2.732 MB/s [ 666.9 IOPS]
Random Read 4KB (QD=32) : 4.655 MB/s [ 1136.5 IOPS]
Random Write 4KB (QD=32) : 2.798 MB/s [ 683.1 IOPS]

Test : 50 MB [E: 0.0% (0.0/14.6 GB)] (x1)
Date : 2014/03/08 11:57:52
OS : Windows 7 Home Premium SP1 [6.1 Build 7601] (x64)

Re: SD Card Benchmarks

Posted: Thu Apr 03, 2014 9:45 pm
by protomouse
Image

Samsung MB-Mpbgca/EU Micro SDHC PLUS Class 10 UHS-I (32 GB)

http://www.amazon.co.uk/dp/B00CNVT442

Re: SD Card Benchmarks

Posted: Wed May 21, 2014 6:46 pm
by davity
Sandisk SDHC Ultra Class 10 UHS-1 8GB
SDSDU-008G-U46

http://www.amazon.es/gp/product/B00812K4V4
-----------------------------------------------------------------------
CrystalDiskMark 3.0.3 x64 (C) 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 20.769 MB/s
Sequential Write : 12.630 MB/s
Random Read 512KB : 20.564 MB/s
Random Write 512KB : 2.115 MB/s
Random Read 4KB (QD=1) : 5.332 MB/s [ 1301.7 IOPS]
Random Write 4KB (QD=1) : 1.348 MB/s [ 329.1 IOPS]
Random Read 4KB (QD=32) : 5.403 MB/s [ 1319.2 IOPS]
Random Write 4KB (QD=32) : 1.090 MB/s [ 266.0 IOPS]

Test : 50 MB [F: 0.8% (63.8/7576.0 MB)] (x5)
Date : 2014/05/21 20:38:19
OS : Windows 7 Professional SP1 [6.1 Build 7601] (x64)
Image

I'm not really sure if it's a UHS-I as Amazon show in the page. I'm amazed about the noobs SD, it's really fast!

Re: SD Card Benchmarks

Posted: Sun May 25, 2014 10:25 pm
by flytomars
Hi all,
My first post here :)
I have been conducting some benchmarking of my own in the last week,
And I got to the conclusion that this is complete voodoo-
2 cards of the same serial number give completely different results,
But each card is consistent with its own results so at least it means the results are valid (even after formatting with different options - results were the same for the specific card).
I have checked 32gb class 4 (which gave good results, actually), 2gb class 2 (which were so so), 16gb class 2 (again, so so ), several 8gb class 6 (which mostly gave OK results, some more than others, sometimes the same serial gave completely different results, but all were poor with the random write), but the best of the bunch was an 8gb class 4!
I have a problem with the last test, mabe someone encountered something similar- during the last test (QD=32) my card hangs and then is not recognized - I suspect it is my USB adapter- this happened with all the faster cards (the ones which had random write higher than 1), so I dont have all results but I am attaching the 8gb class 4 results (sorry, I couldnt find the serial number - is it encoded on the card itself?)

Re: SD Card Benchmarks

Posted: Wed May 28, 2014 4:04 am
by klmnt
Here's the new Samsung Class 6
Samsung Electronics 8GB Micro SDHC with Adapter Upto 24MB/s Class 6 Memory Card (MB-MS08DA/AM)
http://www.amazon.com/Samsung-Electroni ... 453&sr=1-2

Seems not as good as the old class 6 Samsung cards. This one underperformes on sequential ("Upto 24MB/s" ) but 4k seems relatively decent.

Re: SD Card Benchmarks

Posted: Fri Jun 06, 2014 5:59 am
by jethrogb
Samsung 16GB PLUS UHS-1 Class 10 (MB-MPAGCA/AM) - http://www.amazon.com/dp/B00E1XI6N0

Using fio ssd-test.

Summary

Code: Select all

seq-read:      bw=10604  KB/s, iops=2650
rand-read 4K:  bw= 7410.6KB/s, iops=1852
seq-write:     bw= 2859.8KB/s, iops= 714
rand-write 4K: bw= 1343.1KB/s, iops= 335
Details

Code: Select all

seq-read: (g=0): rw=read, bs=4K-4K/4K-4K, ioengine=libaio, iodepth=4
rand-read: (g=1): rw=randread, bs=4K-4K/4K-4K, ioengine=libaio, iodepth=4
seq-write: (g=2): rw=write, bs=4K-4K/4K-4K, ioengine=libaio, iodepth=4
rand-write: (g=3): rw=randwrite, bs=4K-4K/4K-4K, ioengine=libaio, iodepth=4

seq-read: (groupid=0, jobs=1): err= 0: pid=16227
  read : io=636240KB, bw=10604KB/s, iops=2650 , runt= 60002msec
    slat (usec): min=1 , max=3047 , avg=16.16, stdev=11.77
    clat (usec): min=383 , max=13782 , avg=1487.02, stdev=240.75
     lat (usec): min=443 , max=13791 , avg=1503.93, stdev=240.21
    clat percentiles (usec):
     |  1.00th=[  836],  5.00th=[ 1128], 10.00th=[ 1272], 20.00th=[ 1400],
     | 30.00th=[ 1448], 40.00th=[ 1480], 50.00th=[ 1496], 60.00th=[ 1528],
     | 70.00th=[ 1544], 80.00th=[ 1592], 90.00th=[ 1656], 95.00th=[ 1736],
     | 99.00th=[ 2024], 99.50th=[ 2320], 99.90th=[ 3920], 99.95th=[ 4768],
     | 99.99th=[ 8032]
    bw (KB/s)  : min=10264, max=11728, per=100.00%, avg=10608.45, stdev=174.10
    lat (usec) : 500=0.02%, 750=0.14%, 1000=2.59%
    lat (msec) : 2=96.16%, 4=1.01%, 10=0.08%, 20=0.01%
  cpu          : usr=3.37%, sys=6.81%, ctx=114122, majf=0, minf=27
  IO depths    : 1=0.1%, 2=0.1%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued    : total=r=159060/w=0/d=0, short=r=0/w=0/d=0
rand-read: (groupid=1, jobs=1): err= 0: pid=16239
  read : io=444648KB, bw=7410.6KB/s, iops=1852 , runt= 60002msec
    slat (usec): min=2 , max=1150 , avg=16.74, stdev= 8.67
    clat (usec): min=389 , max=13172 , avg=2135.02, stdev=255.90
     lat (usec): min=398 , max=13198 , avg=2152.54, stdev=256.05
    clat percentiles (usec):
     |  1.00th=[ 1576],  5.00th=[ 1768], 10.00th=[ 1864], 20.00th=[ 1960],
     | 30.00th=[ 2024], 40.00th=[ 2096], 50.00th=[ 2160], 60.00th=[ 2192],
     | 70.00th=[ 2224], 80.00th=[ 2320], 90.00th=[ 2384], 95.00th=[ 2448],
     | 99.00th=[ 2800], 99.50th=[ 3024], 99.90th=[ 3952], 99.95th=[ 4704],
     | 99.99th=[ 8256]
    bw (KB/s)  : min= 7144, max= 7768, per=100.00%, avg=7415.26, stdev=74.49
    lat (usec) : 500=0.01%, 750=0.02%, 1000=0.02%
    lat (msec) : 2=26.03%, 4=73.83%, 10=0.09%, 20=0.01%
  cpu          : usr=3.18%, sys=5.71%, ctx=111518, majf=0, minf=27
  IO depths    : 1=0.1%, 2=0.1%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued    : total=r=111162/w=0/d=0, short=r=0/w=0/d=0
seq-write: (groupid=2, jobs=1): err= 0: pid=16241
  write: io=171592KB, bw=2859.8KB/s, iops=714 , runt= 60003msec
    slat (usec): min=1 , max=562 , avg=25.51, stdev=13.52
    clat (usec): min=966 , max=52638 , avg=5560.09, stdev=7613.91
     lat (usec): min=981 , max=52648 , avg=5586.63, stdev=7616.43
    clat percentiles (usec):
     |  1.00th=[ 2064],  5.00th=[ 2512], 10.00th=[ 3056], 20.00th=[ 3152],
     | 30.00th=[ 3216], 40.00th=[ 3248], 50.00th=[ 3312], 60.00th=[ 3376],
     | 70.00th=[ 3888], 80.00th=[ 3984], 90.00th=[ 4128], 95.00th=[31872],
     | 99.00th=[33024], 99.50th=[39680], 99.90th=[48384], 99.95th=[48896],
     | 99.99th=[50944]
    bw (KB/s)  : min=  354, max= 4832, per=100.00%, avg=2859.45, stdev=1329.29
    lat (usec) : 1000=0.01%
    lat (msec) : 2=0.55%, 4=81.07%, 10=9.94%, 20=1.54%, 50=6.87%
    lat (msec) : 100=0.03%
  cpu          : usr=1.72%, sys=3.21%, ctx=31547, majf=0, minf=23
  IO depths    : 1=0.1%, 2=0.1%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued    : total=r=0/w=42898/d=0, short=r=0/w=0/d=0
rand-write: (groupid=3, jobs=1): err= 0: pid=16247
  write: io=80712KB, bw=1343.1KB/s, iops=335 , runt= 60058msec
    slat (usec): min=4 , max=493 , avg=42.25, stdev=14.21
    clat (msec): min=3 , max=500 , avg=11.85, stdev=17.76
     lat (msec): min=3 , max=500 , avg=11.89, stdev=17.76
    clat percentiles (msec):
     |  1.00th=[    6],  5.00th=[    7], 10.00th=[    7], 20.00th=[    8],
     | 30.00th=[    9], 40.00th=[    9], 50.00th=[    9], 60.00th=[   10],
     | 70.00th=[   10], 80.00th=[   10], 90.00th=[   13], 95.00th=[   46],
     | 99.00th=[   60], 99.50th=[  100], 99.90th=[  190], 99.95th=[  388],
     | 99.99th=[  490]
    bw (KB/s)  : min=   60, max= 1944, per=100.00%, avg=1361.77, stdev=668.94
    lat (msec) : 4=0.01%, 10=86.71%, 20=6.76%, 50=4.17%, 100=1.84%
    lat (msec) : 250=0.40%, 500=0.09%, 750=0.01%
  cpu          : usr=1.27%, sys=2.48%, ctx=20355, majf=0, minf=21
  IO depths    : 1=0.1%, 2=0.1%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued    : total=r=0/w=20178/d=0, short=r=0/w=0/d=0

Run status group 0 (all jobs):
   READ: io=636240KB, aggrb=10603KB/s, minb=10603KB/s, maxb=10603KB/s, mint=60002msec, maxt=60002msec

Run status group 1 (all jobs):
   READ: io=444648KB, aggrb=7410KB/s, minb=7410KB/s, maxb=7410KB/s, mint=60002msec, maxt=60002msec

Run status group 2 (all jobs):
  WRITE: io=171592KB, aggrb=2859KB/s, minb=2859KB/s, maxb=2859KB/s, mint=60003msec, maxt=60003msec

Run status group 3 (all jobs):
  WRITE: io=80712KB, aggrb=1343KB/s, minb=1343KB/s, maxb=1343KB/s, mint=60058msec, maxt=60058msec

Disk stats (read/write):
  mmcblk0: ios=224443/51407, merge=45781/11664, ticks=404352/412872, in_queue=817232, util=99.69%

Re: SD Card Benchmarks

Posted: Sat Jul 26, 2014 6:23 pm
by Leopold
Just tested a new card and got some crazy good numbers...

Samsung 16GB Class 10 Grade 1 48MB/s Micro SDHC Plus (MB-MPAGCAEU)
http://www.amazon.co.uk/Samsung-Class-M ... B00D6ENF7M £10.64

-----------------------------------------------------------------------
CrystalDiskMark 3.0.3 x64 (C) 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 46.010 MB/s
Sequential Write : 36.343 MB/s
Random Read 512KB : 45.393 MB/s
Random Write 512KB : 36.277 MB/s
Random Read 4KB (QD=1) : 12.400 MB/s [ 3027.4 IOPS]
Random Write 4KB (QD=1) : 3.445 MB/s [ 841.0 IOPS]
Random Read 4KB (QD=32) : 11.771 MB/s [ 2873.9 IOPS]
Random Write 4KB (QD=32) : 3.591 MB/s [ 876.8 IOPS]

Test : 50 MB [F: 0.0% (0.0/14.6 GB)] (x5)
Date : 2014/07/26 19:10:27
OS : Windows 7 Enterprise SP1 [6.1 Build 7601] (x64)

Re: SD Card Benchmarks

Posted: Sat Jul 26, 2014 6:32 pm
by flytomars
Hi Leopold,
Sorry to say but your test has no meaning as you tested the card not in the pi,
And the pi does not support the benefits of class 10- meaning it would work with the pi, but not give the results you jus got.



Leopold wrote:Just tested a new card and got some crazy good numbers...

Samsung 16GB Class 10 Grade 1 48MB/s Micro SDHC Plus (MB-MPAGCAEU)

-----------------------------------------------------------------------
CrystalDiskMark 3.0.3 x64 (C) 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 46.010 MB/s
Sequential Write : 36.343 MB/s
Random Read 512KB : 45.393 MB/s
Random Write 512KB : 36.277 MB/s
Random Read 4KB (QD=1) : 12.400 MB/s [ 3027.4 IOPS]
Random Write 4KB (QD=1) : 3.445 MB/s [ 841.0 IOPS]
Random Read 4KB (QD=32) : 11.771 MB/s [ 2873.9 IOPS]
Random Write 4KB (QD=32) : 3.591 MB/s [ 876.8 IOPS]

Test : 50 MB [F: 0.0% (0.0/14.6 GB)] (x5)
Date : 2014/07/26 19:10:27
OS : Windows 7 Enterprise SP1 [6.1 Build 7601] (x64)

Re: SD Card Benchmarks

Posted: Sat Jul 26, 2014 6:57 pm
by mikerr
flytomars wrote:Hi Leopold,
Sorry to say but your test has no meaning as you tested the card not in the pi,
And the pi does not support the benefits of class 10- meaning it would work with the pi, but not give the results you jus got.
True, the Pi can't take advantage of the sequential figures, but his 4K random read/write are the important figures for Pi use, and those are high too.
Leopold wrote:Just tested a new card and got some crazy good numbers...

Samsung 16GB Class 10 Grade 1 48MB/s Micro SDHC Plus (MB-MPAGCAEU)
...
Random Write 4KB (QD=32) : 3.591 MB/s [ 876.8 IOPS]