Active Topics

 


Page 2 of 3 1 2 3
Reply
Thread Tools
sixwheeledbeast is online now sixwheeledbeast
2012-10-02 , 19:54
Posts: 2,292 | Thanked: 4,135 times | Joined on Apr 2010 @ UK
#11
The above 64GB card seems to work fine for me.
Has two 800MB swaps and the rest data (FAT).
I don't have any random I/O figures tho.
__________________

Wiki Admin
sixwheeledbeast's wiki
Testing Squad Subscriber
- mcallerx - tenminutecore - FlopSwap - Qnotted - zzztop - Bander - Fight2048 -


Before posting or starting a thread please try this.
 
misterc's Avatar
misterc is offline misterc
2012-10-02 , 19:59
Posts: 1,625 | Thanked: 998 times | Joined on Aug 2010
#12
there you go...
__________________
information is a necessary though no sufficient condition to rationality...
 

The Following User Says Thank You to misterc For This Useful Post:
szopin is offline szopin
2012-10-02 , 20:07
Posts: 2,076 | Thanked: 3,268 times | Joined on Feb 2011
#13
Care to do a stress-test? There are tools/benchmarks to do it, but generating 10000 1kb files and simple >> to them of a single char in random order should help. Random read/writes usually are in orders of 100kb/s, continuous writes for HD recording keep the Mb/s figures from your post.
 
misterc's Avatar
misterc is offline misterc
2012-10-02 , 20:29
Posts: 1,625 | Thanked: 998 times | Joined on Aug 2010
#14
Originally Posted by szopin View Post
Care to do a stress-test? There are tools/benchmarks to do it, but generating 10000 1kb files and simple >> to them of a single char in random order should help. Random read/writes usually are in orders of 100kb/s, continuous writes for HD recording keep the Mb/s figures from your post.
my aim was to get an idea of the performance i may expect from the µSDXC in the N900
(do the tools you refer to run on the N900?)
i guess i could use a dd command with bs=1kb but it wouldn't be random, now, would it?

i have added a 2GB swap partition at the end of the card and reserved most of the space for a ext2 partition for my still to be compiled "local repository" of extras+testing+dev

i guess for swap your kbs size case could apply but otherwise? i don't run an OS on the µSD card, thus...
__________________
information is a necessary though no sufficient condition to rationality...
 
szopin is offline szopin
2012-10-02 , 20:56
Posts: 2,076 | Thanked: 3,268 times | Joined on Feb 2011
#15
There was a tool for benchmarking sd cards straight on maemo, can't find it now... What I found was my post after getting much advertised fast random IO samsung class 10:
http://talk.maemo.org/showpost.php?p...&postcount=132

Benchmark on different OS with usb connected or through card reader shouldn't be that far away though.
 
misterc's Avatar
misterc is offline misterc
2012-10-02 , 21:32
Posts: 1,625 | Thanked: 998 times | Joined on Aug 2010
#16
i'll give the dd with bs=1kb a run for its money next time around, but don't hold your breath, okay?

actually, i also need to put it in the 808 PV and check whether the full resolution pictures save faster on it then on the internal flash (about 5 sec )
considering the sustained write speed & the 10MB size of a 38 MP jpg, that should go down to about 2 sec
unless there is a lot of processing involved
__________________
information is a necessary though no sufficient condition to rationality...
 
misterc's Avatar
misterc is offline misterc
2012-10-02 , 22:41
Posts: 1,625 | Thanked: 998 times | Joined on Aug 2010
#17
Code:
Nokia-N900:/media/mmc1/bonnie# bonnie++ -d . -s 512m -n 1 -b -u user
Using uid:29999, gid:29999.
Writing a byte at a time...done
Writing intelligently...done
Rewriting...done
Reading a byte at a time...done
Reading intelligently...done
start 'em...done...done...done...done...done...
Create files in sequential order...done.
Stat files in sequential order...done.
Delete files in sequential order...done.
Create files in random order...done.
Stat files in random order...done.
Delete files in random order...done.
Version  1.96       ------Sequential Output------ --Sequential Input- --Random-
Concurrency   1     -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
Nokia-N900     512M   254  98  2968   5  3443   5   738  94 15409  13 555.1  47
Latency               139ms    5771ms    2809ms   66437us     199ms     775ms
Version  1.96       ------Sequential Create------ --------Random Create--------
Nokia-N900          -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                  1    82  33 +++++ +++   199  11   134  39 +++++ +++   185  21
Latency               441ms   54198us     104ms   50690us   32500us     102ms
1.96,1.96,Nokia-N900,1,1349212260,512M,,254,98,2968,5,3443,5,738,94,15409,13,555.1,47,1,,,,,82,33,+++++,+++,199,11,134,39,+++++,+++,185,21,139
ms,5771ms,2809ms,66437us,199ms,775ms,441ms,54198us,104ms,50690us,32500us,102ms
to be compared with... this
didn't see that post until after i had started the command so that'll have to do for a 1st run

parameters are not the same, but performance seems roughly 50% higher...
Attached Files
File Type: txt bonnie01.txt (1.5 KB, 61 views)
__________________
information is a necessary though no sufficient condition to rationality...
 

The Following User Says Thank You to misterc For This Useful Post:
szopin is offline szopin
2012-10-02 , 22:55
Posts: 2,076 | Thanked: 3,268 times | Joined on Feb 2011
#18
Please post exact commands to run, will try with my samsung sdhc which gave me x5 boost some half a year ago vs best at that time. Really hoping sd developments are catching up, booting ubuntu might end up being faster than maemo

Or is it just bonnie++?
 
misterc's Avatar
misterc is offline misterc
2012-10-02 , 23:00
Posts: 1,625 | Thanked: 998 times | Joined on Aug 2010
#19
Originally Posted by szopin View Post
Please post exact commands to run, will try with my samsung sdhc which gave me x5 boost some half a year ago vs best at that time. Really hoping sd developments are catching up, booting ubuntu might end up being faster than maemo

Or is it just bonnie++?
hey, don't shoot the messenger, will you?
the command i used is 1st line
on the post i refer to the command is slightly different (lasts over an hr... i had about 10 min)
will get back with later, but might take some time (can't leave the darn thing lying around for an hour not doing anything, now, can i ?!?)

honestly, i think the performances from my 1st post earlier strike me as much more realistic then all that "stat" stuff, but... less "standard" obviously
__________________
information is a necessary though no sufficient condition to rationality...
 
szopin is offline szopin
2012-10-02 , 23:10
Posts: 2,076 | Thanked: 3,268 times | Joined on Feb 2011
#20
Know that feel Seriously interested though. I do remember a thred on tmo comparing random vs consequential io performance of sd cards with some benchmark tool, but my googlefu fails me totally. While looking into it 300kb/s for random on class 10 was around 1/5th (1/4th maybe?) of what fastest class 2 cards were giving and given huge boost in consequential seemed like the best option. Really hope this improved. Even then people were considering moving swap to sd, if it improves like calculating power, we might soon move to keeping rootfs on sd card and get speed boost out of it (not to mention more rootfs space).
 
Reply
Page 2 of 3 1 2 3

« Previous Thread | Next Thread »

 
Forum Jump


All times are GMT. The time now is 16:08.

Contact Us - Home - Archive - Top