[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-usb-users
Subject:    [Linux-usb-users] testing throuput
From:       Carl Karsten <carl () personnelware ! com>
Date:       2007-05-29 14:57:09
Message-ID: 465C3F45.3070100 () personnelware ! com
[Download RAW message or body]

I do some work with PDA style data collection, and one bottleneck is the 
transfer to a PC.  at times there are 100+ units that have to be attached to 
USB, data transfer, repeat.  This can take hours, and often there are 100+ 
people waiting for 'their data' so reducing the time is v.important.

In thinking about this problem, I was wondering how significant the speed of the 
various components are.  my hunch is the USB connection is the slow spot, even 
if it is usb2 (sometimes it is usb1.)  That and the physical handling of the 
equipment.  But I want to test this.

My first thought was: plug in a thumb drive and use dd to blast some data back 
and forth.  except that I got very different results from 2 usb memory sticks 
and a usb/ide HD.  which made me realize that my tests might be flawed, even 
though reading was pretty consistent.

I realize that this may be more effort than it is worth - but every time there 
is a line of cranky people I wonder what things can be done to improve things.

I think what I am looking for is a way to read/write over the USB connection 
that does not actually utilize the storage system on the other side.  kinda like 
what ping does for networking.

Also, has anyone seen better/worse usb hardware?  as in, might a $20 usb 2 card 
be somehow 'worse' than an $80 one?

Below are times for writing/reading 5mb and a baseline to see how long the zero 
and null io took.

Carl K

++ dd if=/dev/zero of=/media/usb/sda1/foo.dat count=5 bs=1M
5+0 records in
5+0 records out
5242880 bytes (5.2 MB) copied, 25.1132 seconds, 209 kB/s

real    0m25.396s
user    0m0.008s
sys     0m0.348s
++ dd if=/dev/zero of=/media/usb/sdb1/foo.dat count=5 bs=1M
5+0 records in
5+0 records out
5242880 bytes (5.2 MB) copied, 119.508 seconds, 43.9 kB/s

real    2m0.147s
user    0m0.004s
sys     0m0.540s
++ dd if=/dev/zero of=/media/usb/sdc1/foo.dat count=5 bs=1M
5+0 records in
5+0 records out
5242880 bytes (5.2 MB) copied, 0.53498 seconds, 9.8 MB/s

real    0m0.554s
user    0m0.004s
sys     0m0.092s
++ dd if=/media/usb/sda1/foo.dat of=/dev/null
10240+0 records in
10240+0 records out
5242880 bytes (5.2 MB) copied, 0.10355 seconds, 50.6 MB/s

real    0m0.115s
user    0m0.020s
sys     0m0.092s
++ dd if=/media/usb/sdb1/foo.dat of=/dev/null
10240+0 records in
10240+0 records out
5242880 bytes (5.2 MB) copied, 0.103117 seconds, 50.8 MB/s

real    0m0.114s
user    0m0.028s
sys     0m0.084s
++ dd if=/media/usb/sdc1/foo.dat of=/dev/null
10240+0 records in
10240+0 records out
5242880 bytes (5.2 MB) copied, 0.103213 seconds, 50.8 MB/s

real    0m0.114s
user    0m0.016s
sys     0m0.100s


root@dhcp120:~# time dd if=/dev/zero of=/dev/null count=10 bs=1M
10485760 bytes (10 MB) copied, 0.00134113 seconds, 7.8 GB/s
real    0m0.014s
user    0m0.008s
sys     0m0.004s


-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Linux-usb-users@lists.sourceforge.net
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-users
[prev in list] [next in list] [prev in thread] [next in thread]