Name: SAMSUNG EVO UHS-1 HC-1 CL10
Size: 16GB ( 15.7GB, 15720251392 bytes )
Allocation Unit: 4MB
Write Size: 32KB
Page Size: 8KB
FAT Location: -
# open Aus linear: 28
# open Aus random: 28
Algorithm: random sector or page access ?
Manuf ID: 0x1B
OEMID: 0x534D
NOTE: Another card with Manuf ID: 0x1D showed very inferior random write
speeds compared to this one ( about 2x less ). I suspect it may be a fake,
and curiously enough the good one was bought from Hong Kong on Ebay while
this other was bought on a major store in my country.
Hi there,
Here are my results for the SanDisk Cruzer Blade 16GB tests. The unit seems
to behave a bit better than the smaller ones already in the list. The USB
ID is identical to those of the smaller ones, however.
At first, flashbench -a gave an uneven list of diff values. After reading
that amount of writes can make a difference, I zeroed the (brand new) drive
to all zeroes: dd if=/dev/zero of=/dev/sdx bs=4M. I then got the following:
sudo flashbench -a /dev/sdf --blocksize=1024
align 4294967296 pre 748µs on 1.09ms post 756µs
diff 335µs
align 2147483648 pre 749µs on 1.16ms post 751µs
diff 406µs
align 1073741824 pre 748µs on 1.15ms post 751µs
diff 402µs
align 536870912 pre 746µs on 1.15ms post 750µs diff 401µs
align 268435456 pre 752µs on 1.16ms post 752µs diff 408µs
align 134217728 pre 738µs on 1.15ms post 756µs diff 406µs
align 67108864 pre 749µs on 1.14ms post 749µs diff 395µs
align 33554432 pre 753µs on 1.14ms post 752µs diff 392µs
align 16777216 pre 746µs on 1.14ms post 754µs diff 394µs
align 8388608 pre 745µs on 1.13ms post 747µs diff 388µs
align 4194304 pre 744µs on 983µs post 747µs diff 237µs
align 2097152 pre 752µs on 885µs post 744µs diff 137µs
align 1048576 pre 753µs on 893µs post 753µs diff 140µs
align 524288 pre 752µs on 880µs post 739µs diff 134µs
align 262144 pre 750µs on 891µs post 757µs diff 137µs
align 65536 pre 748µs on 888µs post 750µs diff 139µs
align 32768 pre 749µs on 883µs post 745µs diff 137µs
align 16384 pre 750µs on 890µs post 751µs diff 139µs
align 8192 pre 745µs on 803µs post 748µs diff 56.8µs
align 4096 pre 747µs on 804µs post 753µs diff 53.5µs
align 2048 pre 746µs on 803µs post 745µs diff 57.3µs
Leading to an assumption of erase size of 4M and page size of 8K.
With a sequence of
sudo flashbench -O --erasesize=$[4*1024*1024] --blocksize=$[8*1024]
/dev/sdf --open-au-nr=X
Where X increased linearly from 2 to 10 I determined the cutoff at AU=8
Because of:
AU=2
4MiB 22M/s
2MiB 21.9M/s
1MiB 21.9M/s
512KiB 21.6M/s
256KiB 20.9M/s
128KiB 19.3M/s
64KiB 18.7M/s
32KiB 16.2M/s
16KiB 11.4M/s
8KiB 6.99M/s
AU=7
4MiB 22M/s
2MiB 21.9M/s
1MiB 21.8M/s
512KiB 21.6M/s
256KiB 20.6M/s
128KiB 19.1M/s
64KiB 18.3M/s
32KiB 15.7M/s
16KiB 10.8M/s
8KiB 6.58M/s
AU=8
4MiB 14.4M/s
2MiB 21.4M/s
1MiB 21.8M/s
512KiB 21.6M/s
256KiB 20.6M/s
128KiB 19.1M/s
64KiB 18.2M/s
32KiB 15.6M/s
16KiB 10.7M/s
8KiB 6.47M/s
AU=9
4MiB 11.3M/s
2MiB 7M/s
1MiB 4.58M/s
512KiB 5.15M/s
256KiB 5.4M/s
128KiB 4.96M/s
64KiB 5.84M/s
32KiB 5.17M/s
16KiB 4.91M/s
8KiB 3.7M/s
Random performance sucks at smaller sizes. Results nearly the same for AU=1
through 6
sudo flashbench -O --erasesize=$[4*1024*1024] --blocksize=$[8*1024]
/dev/sdf --random --open-au-nr=6
[sudo] password for gaby:
4MiB 22.2M/s
2MiB 22M/s
1MiB 22M/s
512KiB 20M/s
256KiB 9.8M/s
128KiB 2.69M/s
64KiB 1.12M/s
32KiB 563K/s
16KiB 285K/s
8KiB 132K/s
If there's anything else I can test/check, let me know?
Regards,
Gaby Schilders
I am not subscribed to linaro-dev, so please Cc me in case you drop flashbench-results
mailinglist from receivers.
Hi Arnd, hi everyone!
For now quick and short.
A simple script to automate common steps on reporting flash medium test
results. It goes as far as read tests for erase block size and then suggests
some open au tests together with a warning that these write to the device.
I checked other flashbench-results posts as to what info would be
interesting.
Find it at:
git://gitorious.org/flashbench/flashbench.git
Example output is at:
[Flashbench] Samsung Plus MB-SPAGFFP Class 10 SDHC 16GB
Sat Dec 1 16:02:18 UTC 2012
http://lists.linaro.org/pipermail/flashbench-results/2012-December/000346.h…
Sorry that I occupied flashbench upstream name for project. I can look
into renaming it to "flashbench-martin" or something like that.
The script is in branch "flashreport".
In branch "defaultblocksize" you find a currently untested patch for
changing default block size to 1024.
I can try to send patches as mails via git, but I´d have to look it up first,
since I do this quite rarely.
Up to then please use the repo urls :)
Nice weekend,
--
Martin 'Helios' Steigerwald - http://www.Lichtvoll.de
GPG: 03B0 0D6C 0040 0710 4AFA B82F 991B EAAC A599 84C7
Salut,
for housing a debian chroot in my android phone, I want to set up a 2.5G
partition on this new micro sd card. For that, I want to figure out
which parameters to give to mke2fs, as in
> mkfs.ext4 -E stride=2,stripe-width=2048 -L debian_chroot /dev/sdb2
Now to figure out the internal erase block size, flashbench was
recommended. But I can not quite make sense of my result..
> # fdisk.distrib -l /dev/sdb
>
> Disk /dev/sdb: 15.9 GB, 15931539456 bytes
> 255 heads, 63 sectors/track, 1936 cylinders, total 31116288 sectors
> Units = sectors of 1 * 512 = 512 bytes
> Sector size (logical/physical): 512 bytes / 512 bytes
> I/O size (minimum/optimal): 512 bytes / 512 bytes
> Disk identifier: 0x000604f9
>
> Device Boot Start End Blocks Id System
> /dev/sdb1 2048 25890815 12944384 b W95 FAT32
> /dev/sdb2 25890816 31115263 2612224 83 Linux
> # flashbench -va /dev/sdb --blocksize=1024 |(echo "align,pre [µs],on [µs],post [µs],diff [µs]" && sed -r 's/\w+ //g;s/\t/,/g;s/[0-9]+ns*/0/g;s/µs//g')|tee flashbench-16gb-sandisk-extreme-microsdhc-blocksize-1024.csv
> # flashbench -va /dev/sdb --blocksize=$[3*1024] |(echo "align,pre [µs],on [µs],post [µs],diff [µs]" && sed -r 's/\w+ //g;s/\t/,/g;s/[0-9]+ns*/0/g;s/µs//g')|tee flashbench-16gb-sandisk-extreme-microsdhc-blocksize-$[3*1024].csv
resulting CSV files attached.. I created a spreadsheet chart showing a
drop.. does this mean the erasure block size is 128KB?
That would give a stripe-width of 128KB/(2*4KB)=16?
And I have not at all grokked the benefit/applicability of
open-au/open-au-nr for practical purposes..
Any clarification greatly appreciated.
#Regards!Marcel