Transfer rates of external USB 2.0 hard drive
Joe Negron
Amazon.com). I'm curious if the transfer rates I'm getting are about
right.
I copied 61.08 GB in 7,915 files (~7.90 MB/file) requiring 159:45
(MM:SS), yielding a transfer rate of 6,682 KB/sec from a JFS partition
on the internal drive to a JFS partition on the external drive (I
partitioned the entire external drive as one 500 GB volume).
I'm confused because the SimpleTech documentation states that a USB 2.0
connection could yield transfer rates "up to 480 Mb/sec" (and 12 Mb/sec
for USB 1.1). Note the lower-case "b" in "Mb/sec" - I don't know if they
mean bits or bytes. If they mean bits and "M" means 1,000,000, then:
480,000,000 / 8 = 60,000,000 = 58,593.75 KB/sec
which is much faster than my results.
If they mean bits and "M" means 1024 * 1024, then:
480 * 1,048,576 / 8 = 62,914,560 = 61,440 KB/sec
And, if they mean megabytes (not megabits) then their claim is even more
out of line with my results:
480 * 1,048,576 = 503,316,480 = 491,520 KB/sec
Note that I first did some trials with Windows XP. While I didn't copy
near the number or size of files as I did with OS/2 and JFS, I got
similar speeds with NTFS->NTFS, and signficantly slower speeds with
NTFS->FAT32.
So, this isn't an OS/2 issue.
Of course I understand their claim is likely of an ideal, but my results
are orders of magnitude slower.
--
------------------------------------------------------------------------
If you steal something small you are a petty thief, but if you steal
millions you are a gentleman of society.
--Greek Proverb
War is good for business - invest your son.
--antiwar bumper sticker from the 1960s
------------------------------------------------------------------------
Joe Negron from Sheepshead Bay, Brooklyn, NY, USA
--
Posted via a free Usenet account from http://www.teranews.com
Rodney Pont
>I recently purchased a SimpleTech 500 GB USB 2.0 External HD ($169.99 at
>Amazon.com). I'm curious if the transfer rates I'm getting are about
>right.
>
>I copied 61.08 GB in 7,915 files (~7.90 MB/file) requiring 159:45
>(MM:SS), yielding a transfer rate of 6,682 KB/sec from a JFS partition
>on the internal drive to a JFS partition on the external drive (I
>partitioned the entire external drive as one 500 GB volume).
It's the number of files, there is a lot of work to be done for each
file so you are getting lots of pauses while it updates the
directories. Try with a large file and see how fast it goes.
--
Regards - Rodney Pont
The from address exists but is mostly dumped,
please send any emails to the address below
e-mail ngpsm4 (at) infohitsystems (dot) ltd (dot) uk
Peter Brown
I've just been rearranging my partitions with the help of a 40Gb USB drive.
I have to say that I did not think the copy speed from internal (either
hpfs or jfs format) to external (jfs format) drive was that fast but I
seem to have had better luck than you going by the few timed results
(not using a stop watch :-) that I have:-
Copying 6,366 files, 1,533,688Kb (1497Mb approx) took 10 mins 35 secs -
give or take a sec or 2. This was internal hpfs to external jfs.
That seems to indicate a write speed of 1533688/635 secs =
2415.256693Kb/sec - approx 2.35Mb/sec
Another result was 3752Mb written in 18 mins 5 secs - only 2 files
involved here, both files being my VPC win2000 drive. This was internal
jfs to external jfs.
I suspect that Rodney may have a point about the number of files involved...
I did notice when copying files back to resized partitions that the copy
speed seemed to be a bit quicker from USB to internal (SCSI U160) drive.
Having looked at this copying with the USB drive initially formatted
with fat32 I have to state that jfs is about twice as fast in both
directions.
Regards
Pete
Joe Negron
> On 16 Mar 2007 11:02:16 GMT, Joe Negron wrote:
>
>>I recently purchased a SimpleTech 500 GB USB 2.0 External HD ($169.99 at
>>Amazon.com). I'm curious if the transfer rates I'm getting are about
>>right.
>>
>>I copied 61.08 GB in 7,915 files (~7.90 MB/file) requiring 159:45
>>(MM:SS), yielding a transfer rate of 6,682 KB/sec from a JFS partition
>>on the internal drive to a JFS partition on the external drive (I
>>partitioned the entire external drive as one 500 GB volume).
>
> It's the number of files, there is a lot of work to be done for each
> file so you are getting lots of pauses while it updates the
> directories. Try with a large file and see how fast it goes.
Well, you're partially correct. I re-ran my tests, this time with 114
files, each 50 MB or larger. There were a total of 114 files, totalling
9.67 GB (~86.89 MB/file). It required 16:37 (MM:SS) to complete the
copy, yielding a transfer rate of 10,177 KB/sec, which is still way less
than the minimum transfer rate specified by the manufacturer SimpleTech
(cf. the confusion of whether "M" means million or mega, and whether "b"
means bits or bytes), anywhere from 58,593.75 KB/sec (million, bits) to
61,440 KB/sec (mega, bits), to 491,520 KB/sec (mega, bytes).
So, I'm wondering whether or not the documentation is incorrect, there
is something wrong with my setup (if so, it should be a hardware issue
since I get similar speeds with Windowx XP), or if there's something
wrong with the drive.
--
------------------------------------------------------------------------
No one can make you feel inferior without your consent.
--Eleanor Roosevelt
dinkmeister
On 17 Mar 2007 03:24:55 GMT, Joe Negron wrote:
:
Ilya Zakharevich
dinkmeister
<di...@yadda.com>], who wrote in article <qvaxlnqqnpbz.j...@newsgroups.comcast.net>:
> USB is just slow as hell on os2, unfortunately.
Do not think so. My JFS-through-USB1 writes at 900K/sec (with large
files). FAT32 on WinXT writes at about 950K/sec. Looks sane...
(Of course, the situation with USB2 may be very different - but I do
not have it.)
The tests the OP did are just, IMO, very flawed. With such
ridiculously small files (50M!!!) the write speed would have no
relationship to the disk throughput.
Hope this helps,
Ilya
Heiko Nitzsche
limit of the bus when doing bidirectional transfers. So half the
speed for unidirectional operation and you get 240 Mbit/sec for
one direction.
As we're talking about bits per second here, we get a theoretical
transfer rate in one direction of 30MByte/sec (240 Mbit/sec / 8).
As there is some protocol overhead for maintaining the USB connection,
you will never get more than 29-29.5MByte/sec with USB 2.0 Highspeed.
This is also what chipset tests show.
My external USB 2.0 Highspeed disk (300GB TrekStor) achieves a
realistic transfer rate of 27-28MByte/sec under Win XP with
NTFS->NTFS and large files.
On eCS with the latest USB drivers the transfer rate is only
marginally less than this, it's around 27MByte/sec with JFS.
With FAT32 on eCS it's very, very slow, less than 2MByte/sec.
So this is useless for transferring much data.
As others noted already, the overhead increases with the file/size
ratio. The smaller the files the higher the overhead compared to
the file data.
So it's not the fault of the drivers or the hardware.
If you get much lower transfer rates, check that in BIOS the
USB 2.0 Highspeed mode is enabled. Also try to connect the drive
to an other USB 2.0 port. To get the highest speed, you should
ensure that this port is not shared with other devices (via hub
for instance). Otherwise the bandwidth is shared between these devices.
In Win XP you can see the devices using the same physical USB port
in the Hardware Manager (open the USB port icons). I thin on OS/2
the USB Resource Manager also shows the physical port of devices.
Hope that helps.
Ilya Zakharevich
<hn-expire...@arcor.de>], who wrote in article <45fbeca0$0$6398$9b4e...@newsspool2.arcor-online.net>:
> As we're talking about bits per second here, we get a theoretical
> transfer rate in one direction of 30MByte/sec (240 Mbit/sec / 8).
Can't be so. People measure 33MB/sec transfers in real-life
situations; see http://www.barefeats.com/usb2.html.
But there is some major flaw in USB2 design-or-implemenation - "about
1/2 of the maximum" is what people typically observe...
> marginally less than this, it's around 27MByte/sec with JFS.
> With FAT32 on eCS it's very, very slow, less than 2MByte/sec.
My experience of efficiency of os2's FAT32 is about 6% of the channel
throughput. So it supports this number.
> So this is useless for transferring much data.
"Useless" is a very strong word. I participated in a transfer of more
than 200GB of data through USB1 to a FAT32 drive. A couple of weeks,
and it is there! ;-)
Hope this helps,
Ilya
Heiko Nitzsche
>> transfer rate in one direction of 30MByte/sec (240 Mbit/sec / 8).
I seem to somehow mixed this up with the WLAN 54Mbps promises.
> But there is some major flaw in USB2 design-or-implemenation - "about
> 1/2 of the maximum" is what people typically observe...
I had a quick look at the USB 2.0 specification (see http://www.usb.org)
and found the following interesting note about bulk transfers:
"All Host Controllers are required to have support for 8-, 16-, 32-,
and 64-byte maximum packet sizes for full-speed bulk endpoints and
512 bytes for high-speed bulk endpoints. No Host Controller is required
to support larger or smaller maximum packet sizes.
During configuration, the USB System Software reads the endpoint’s
maximum data payload size and ensures that no data payload will be
sent to the endpoint that is larger than the supported size."
Below this note is a table showing the achievable data rates on
page 55: (overhead for protocol included)
Data Max Bandwidth
Payload (bytes/second)
1 1064000
2 2096000
4 4064000
8 7616000
16 13440000
32 22016000
64 32256000 <-- Is is that what we see?
128 40960000
256 49152000
512 53248000
So the reason for the low transfer rates for USB hard drives
seems to be a manufacturer made restriction in the device USB
interface controller. They seem to support only payload packets
of up to 64 bytes while the mainboard chipset controllers and
the drives could do faster.
Maybe the device manufacturers just try to save the money for
faster interface controllers?
>> So this is useless for transferring much data.
>
> "Useless" is a very strong word. I participated in a transfer of more
> than 200GB of data through USB1 to a FAT32 drive. A couple of weeks,
> and it is there! ;-)
Yeah, if you accept spending that much time ;)
Kevin K
By the time you take into account file system overhead, operating
system overhead, hard drive latencies, that is probably in the
ballpark of what you can expect. It is definately faster than USB
1.1. I don't recall the speed for my last write to an USB drive,
though.
Joe Negron
>>
>>I copied 61.08 GB in 7,915 files (~7.90 MB/file) requiring 159:45
>>(MM:SS), yielding a transfer rate of 6,682 KB/sec from a JFS partition
>>on the internal drive to a JFS partition on the external drive (I
>>partitioned the entire external drive as one 500 GB volume).
>>
>>I'm confused because the SimpleTech documentation states that a USB 2.0
>>connection could yield transfer rates "up to 480 Mb/sec" (and 12 Mb/sec
>>for USB 1.1). Note the lower-case "b" in "Mb/sec" - I don't know if they
>>mean bits or bytes. If they mean bits and "M" means 1,000,000, then:
>>
>> 480,000,000 / 8 = 60,000,000 = 58,593.75 KB/sec
>>
>>which is much faster than my results.
>>
>>If they mean bits and "M" means 1024 * 1024, then:
>>
>> 480 * 1,048,576 / 8 = 62,914,560 = 61,440 KB/sec
>>
>>And, if they mean megabytes (not megabits) then their claim is even more
>>out of line with my results:
>>
>> 480 * 1,048,576 = 503,316,480 = 491,520 KB/sec
>>
>I've just been rearranging my partitions with the help of a 40Gb USB drive.
That's part of the reason I bought this drive. :)
>[...]
>
>Copying 6,366 files, 1,533,688Kb (1497Mb approx) took 10 mins 35 secs -
>give or take a sec or 2. This was internal hpfs to external jfs.
>
>That seems to indicate a write speed of 1533688/635 secs =
>2415.256693Kb/sec - approx 2.35Mb/sec
>
>Another result was 3752Mb written in 18 mins 5 secs - only 2 files
>involved here, both files being my VPC win2000 drive. This was internal
>jfs to external jfs.
So, in your first test your average filesize was a mere 240.92 KB while
in my first test the average filesize was much larger at 7.90 MB,
indicating that the problem I'm having is not very related to overhead.
>I suspect that Rodney may have a point about the number of files involved...
As you should have seen by now, I re-tested with a much larger average
filesize (86.89 MB) and while I got better results, the transfer rate
is still much slower than it seems it should be.
>I did notice when copying files back to resized partitions that the copy
>speed seemed to be a bit quicker from USB to internal (SCSI U160) drive.
Perhaps your internal drive has better write speed than the USB drive .
--
------------------------------------------------------------------------
Imagination is more important than knowledge.
--Albert Einstein
War is good for business - invest your son.
--antiwar bumper sticker from the 1960s
------------------------------------------------------------------------
Joe Negron from Sheepshead Bay, Brooklyn, NY, USA
--
Joe Negron
>USB is just slow as hell on os2, unfortunately.
Perhaps you missed the part where I said I got similar speeds under
Windows XP.
--
------------------------------------------------------------------------
One meets his destiny often in the road he takes to avoid it.
--French Proverb
Joe Negron
>[...]
>The tests the OP did are just, IMO, very flawed. With such
>ridiculously small files (50M!!!) the write speed would have no
>relationship to the disk throughput.
Have you seen the results of my second test? If so, what have you to
say?
--
------------------------------------------------------------------------
If I am not for myself, who will be for me? And if I am only for
myself, what am I? And if not now--when?
--Hillel
Ilya Zakharevich
<joen...@XmindspringX.com>], who wrote in article <45fcf139$0$16321$8826...@free.teranews.com>:
> >The tests the OP did are just, IMO, very flawed. With such
> >ridiculously small files (50M!!!) the write speed would have no
> >relationship to the disk throughput.
> Have you seen the results of my second test? If so, what have you to
> say?
As it is easy to see, I was commenting on your second test. My
experience is that one needs about 1GB files to observe saturation of
the transfer speed.
Hope this helps,
Ilya
Joe Negron
> [A complimentary Cc of this posting was sent to
> Joe Negron
No it wasn't.
><joen...@XmindspringX.com>], who wrote in article <45fcf139$0$16321$8826...@free.teranews.com>:
>>>The tests the OP did are just, IMO, very flawed. With such
>>>ridiculously small files (50M!!!) the write speed would have no
>>>relationship to the disk throughput.
>>Have you seen the results of my second test? If so, what have you to
>>say?
>As it is easy to see, I was commenting on your second test. My
>experience is that one needs about 1GB files to observe saturation of
>the transfer speed.
<sigh>
OK, this time I copied one file, 1,513,355,264 bytes (~1.41 GB). It is
the largest file on my system. I also timed copying the same file to
another USB 2.0 device: a Corsair Voyager 4 GB flash drive formatted as
a single HPFS partition for its entire capacity.
Also included are the previous two tests:
Elapsed Bytes Filesize Xfer rate
(secs) (GB) Files avg. (MB) (MB/sec.)
SimpleDrive (JFS->JFS) 9,585 61.08 7,915 7.90 6.53
SimpleDrive (JFS->JFS) 997 9.67 114 86.86 9.93
SimpleDrive (JFS->JFS) 210 1.41 1 1,443.84 6.87
Voyager (JFS->HPFS) 190 1.41 1 1,443.84 7.61
Can you explain this? Based on the other posts it seems there is
something wrong with my setup, though what that could be I don't know.
There doesn't seem to be anything wrong with the drive since I get a
similar rate copying to a Flash drive.
>Hope this helps,
Not yet, but perhaps in the future. :)
--
------------------------------------------------------------------------
It is left, therefore, to the juries, if they think the permanent judges
are under any bias whatever in any cause, to take on themselves to judge
the law as well as the fact. They never exercise this power but when
they suspect partiality in the judges, and by the exercise of this power
they have been the firmest bulwarks of English liberty.
--Thomas Jefferson
Joe Negron
Thanks for trying to clear this up.
>Forget about the marketing numbers.
Yes, I figured at least some of the differential was the result of
marketing hype.
>This is the theoretical speed
>limit of the bus when doing bidirectional transfers. So half the
>speed for unidirectional operation and you get 240 Mbit/sec for
>one direction.
>
>As we're talking about bits per second here, we get a theoretical
>transfer rate in one direction of 30MByte/sec (240 Mbit/sec / 8).
>As there is some protocol overhead for maintaining the USB connection,
>you will never get more than 29-29.5MByte/sec with USB 2.0 Highspeed.
>This is also what chipset tests show.
>
>My external USB 2.0 Highspeed disk (300GB TrekStor) achieves a
>realistic transfer rate of 27-28MByte/sec under Win XP with
>NTFS->NTFS and large files.
The transfer rates I'm getting are nowhere close to this.
>On eCS with the latest USB drivers
Which are? I have:
12-21-05 14:13 23,386 usbd.sys
12-21-05 14:14 42,112 usbehcd.sys
12-21-05 14:14 41,984 usbmsd.add
12-21-05 14:14 35,532 usbohcd.sys
7-10-06 4:28 32,326 usbresmg.sys
In case it matters, follows are the USB related entries in CONFIG.SYS:
BASEDEV=usbehcd.sys /V
BASEDEV=usbohcd.sys /V
BASEDEV=usbohcd.sys /V
BASEDEV=usbd.sys
BASEDEV=usbmsd.add /V /FLOPPIES:0 /REMOVABLES:3
DEVICE=C:\OS2\BOOT\usbresmg.sys /V
>the transfer rate is only
>marginally less than this, it's around 27MByte/sec with JFS.
>With FAT32 on eCS it's very, very slow, less than 2MByte/sec.
>So this is useless for transferring much data.
Yes, based on a few cursory tests I came to the same conclusion. My
original plan was to have some of these files available when booting XP
but it would take too long to put them there!
>As others noted already, the overhead increases with the file/size
>ratio. The smaller the files the higher the overhead compared to
>the file data.
Please see my recent message to Ilya - while increasing the filesize
increases the transfer rate I'm still not getting close to the numbers I
should be getting (whether under eCS or XP).
>So it's not the fault of the drivers or the hardware.
? But, it sounds like there's something wrong with the hardware, or at
least the configuration of the hardware (though I don't know what I can
do to configure it differently).
>If you get much lower transfer rates, check that in BIOS the
>USB 2.0 Highspeed mode is enabled.
There exists no option in my BIOS with that precise designation. There
is a "USB Controllers" setting, with the following options: "V1.1+V2.0",
"V1.1", and "NONE". Of course, I have it set to the first of these.
>Also try to connect the drive
>to an other USB 2.0 port. To get the highest speed, you should
>ensure that this port is not shared with other devices (via hub
>for instance). Otherwise the bandwidth is shared between these devices.
>In Win XP you can see the devices using the same physical USB port
>in the Hardware Manager (open the USB port icons). I thin on OS/2
>the USB Resource Manager also shows the physical port of devices.
Yes, I've tried this with no difference.
--
------------------------------------------------------------------------
Every man has a right to be wrong in his opinions. But no man has a
right to be wrong in his facts.
--Bernard Baruch
Ilya Zakharevich
Joe Negron
> Elapsed Bytes Filesize Xfer rate
> (secs) (GB) Files avg. (MB) (MB/sec.)
>
> SimpleDrive (JFS->JFS) 9,585 61.08 7,915 7.90 6.53
> SimpleDrive (JFS->JFS) 997 9.67 114 86.86 9.93
> SimpleDrive (JFS->JFS) 210 1.41 1 1,443.84 6.87
> Voyager (JFS->HPFS) 190 1.41 1 1,443.84 7.61
>
> Can you explain this?
"This" being *what*? Large file giving a smaller transfer speed?
I do not observe this on my system (copying done using File
Commander). All I can say is that this should *extremely* sensitive
to the strategy used by copying program: does it preallocate the space
for the target? What size of chunks it uses from writing? Does it
try to interleave reading and writing (multithreaded copy)?
Sorry that I can't be of more help,
Ilya
Joe Negron
Thanks for trying to clear this up.
>Forget about the marketing numbers.
Yes, I figured at least some of the differential was the result of
marketing hype.
>This is the theoretical speed
>limit of the bus when doing bidirectional transfers. So half the
>speed for unidirectional operation and you get 240 Mbit/sec for
>one direction.
>
>As we're talking about bits per second here, we get a theoretical
>transfer rate in one direction of 30MByte/sec (240 Mbit/sec / 8).
>As there is some protocol overhead for maintaining the USB connection,
>you will never get more than 29-29.5MByte/sec with USB 2.0 Highspeed.
>This is also what chipset tests show.
>
>My external USB 2.0 Highspeed disk (300GB TrekStor) achieves a
>realistic transfer rate of 27-28MByte/sec under Win XP with
>NTFS->NTFS and large files.
The transfer rates I'm getting are nowhere close to this.
>On eCS with the latest USB drivers
Which are? I have:
12-21-05 14:13 23,386 usbd.sys
12-21-05 14:14 42,112 usbehcd.sys
12-21-05 14:14 41,984 usbmsd.add
12-21-05 14:14 35,532 usbohcd.sys
7-10-06 4:28 32,326 usbresmg.sys
In case it matters, follows are the USB related entries in CONFIG.SYS:
BASEDEV=usbehcd.sys /V
BASEDEV=usbohcd.sys /V
BASEDEV=usbohcd.sys /V
BASEDEV=usbd.sys
BASEDEV=usbmsd.add /V /FLOPPIES:0 /REMOVABLES:3
DEVICE=C:\OS2\BOOT\usbresmg.sys /V
>the transfer rate is only
>marginally less than this, it's around 27MByte/sec with JFS.
>With FAT32 on eCS it's very, very slow, less than 2MByte/sec.
>So this is useless for transferring much data.
Yes, based on a few cursory tests I came to the same conclusion. My
original plan was to have some of these files available when booting XP
but it would take too long to put them there!
>As others noted already, the overhead increases with the file/size
>ratio. The smaller the files the higher the overhead compared to
>the file data.
Please see my recent message to Ilya - while increasing the filesize
increases the transfer rate I'm still not getting close to the numbers I
should be getting (whether under eCS or XP).
>So it's not the fault of the drivers or the hardware.
? But, it sounds like there's something wrong with the hardware, or at
least the configuration of the hardware (though I don't know what I can
do to configure it differently).
>If you get much lower transfer rates, check that in BIOS the
>USB 2.0 Highspeed mode is enabled.
There exists no option in my BIOS with that precise designation. There
is a "USB Controllers" setting, with the following options: "V1.1+V2.0",
"V1.1", and "NONE". Of course, I have it set to the first of these.
>Also try to connect the drive
>to an other USB 2.0 port. To get the highest speed, you should
>ensure that this port is not shared with other devices (via hub
>for instance). Otherwise the bandwidth is shared between these devices.
>In Win XP you can see the devices using the same physical USB port
>in the Hardware Manager (open the USB port icons). I thin on OS/2
>the USB Resource Manager also shows the physical port of devices.
Yes, I've tried this with no difference.
--
------------------------------------------------------------------------
Every man has a right to be wrong in his opinions. But no man has a
right to be wrong in his facts.
--Bernard Baruch
War is good for business - invest your son.
Joe Negron
> Joe Negron
No it wasn't.
><joen...@XmindspringX.com>], who wrote in article <45fcf139$0$16321$8826...@free.teranews.com>:
>>>The tests the OP did are just, IMO, very flawed. With such
>>>ridiculously small files (50M!!!) the write speed would have no
>>>relationship to the disk throughput.
>>Have you seen the results of my second test? If so, what have you to
>>say?
>As it is easy to see, I was commenting on your second test. My
>experience is that one needs about 1GB files to observe saturation of
>the transfer speed.
<sigh>
OK, this time I copied one file, 1,513,355,264 bytes (~1.41 GB). It is
the largest file on my system. I also timed copying the same file to
another USB 2.0 device: a Corsair Voyager 4 GB flash drive formatted as
a single HPFS partition for its entire capacity.
Also included are the previous two tests:
Elapsed Bytes Filesize Xfer rate
(secs) (GB) Files avg. (MB) (MB/sec.)
SimpleDrive (JFS->JFS) 9,585 61.08 7,915 7.90 6.53
SimpleDrive (JFS->JFS) 997 9.67 114 86.86 9.93
SimpleDrive (JFS->JFS) 210 1.41 1 1,443.84 6.87
Voyager (JFS->HPFS) 190 1.41 1 1,443.84 7.61
Can you explain this? Based on the other posts it seems there is
something wrong with my setup, though what that could be I don't know.
There doesn't seem to be anything wrong with the drive since I get a
similar rate copying to a Flash drive.
>Hope this helps,
Not yet, but perhaps in the future. :)
--
------------------------------------------------------------------------
It is left, therefore, to the juries, if they think the permanent judges
are under any bias whatever in any cause, to take on themselves to judge
the law as well as the fact. They never exercise this power but when
they suspect partiality in the judges, and by the exercise of this power
they have been the firmest bulwarks of English liberty.
--Thomas Jefferson
War is good for business - invest your son.
Heiko Nitzsche
>
> Which are? I have:
> 12-21-05 14:13 23,386 usbd.sys
> 12-21-05 14:14 42,112 usbehcd.sys
> 12-21-05 14:14 41,984 usbmsd.add
> 12-21-05 14:14 35,532 usbohcd.sys
> 7-10-06 4:28 32,326 usbresmg.sys
Yeah, based on the file sizes yours seem to be the same ones.
bldlevel usbd.sys -> 10.162
bldlevel usbehci.sys -> 10.162
bldlevel usbuhcd.sys -> 10.162
bldlevel usbmsd.add -> 10.162
bldlevel os2dasd.dmd -> 14.104
> In case it matters, follows are the USB related entries in CONFIG.SYS:
>
> BASEDEV=usbehcd.sys /V
> BASEDEV=usbohcd.sys /V
> BASEDEV=usbohcd.sys /V
> BASEDEV=usbd.sys
> BASEDEV=usbmsd.add /V /FLOPPIES:0 /REMOVABLES:3
> DEVICE=C:\OS2\BOOT\usbresmg.sys /V
I have a VIA K8T800 chipset with 8237 southbridge (Athlon64)
and thus use the UHCI driver.
Some older nVidia chipsets were known to have lousy USB throuput
but I can't remember which ones.
>> As others noted already, the overhead increases with the file/size
>> ratio. The smaller the files the higher the overhead compared to
>> the file data.
>
> Please see my recent message to Ilya - while increasing the filesize
> increases the transfer rate I'm still not getting close to the numbers I
> should be getting (whether under eCS or XP).
Well, this really sounds as a hardware issue as also your
XP makes it not faster.
>> If you get much lower transfer rates, check that in BIOS the
>> USB 2.0 Highspeed mode is enabled.
>
> There exists no option in my BIOS with that precise designation. There
> is a "USB Controllers" setting, with the following options: "V1.1+V2.0",
> "V1.1", and "NONE". Of course, I have it set to the first of these.
What chipset is this?
Are you sure that it really supports USB 2.0 Highspeed mode
and not just Fullspeed? Simply "USB 2.0" says nothing about the
supported transfer rate. Usually Highspeed capable devices and
controllers have a separate logo (USB 2.0 Highspeed).
If the board only supports Fullspeed mode (12MBit/sec max.) you can
try with an add-on USB card that supports USB 2.0 Highspeed mode.
Even though I haven't done this, the USB driver readme says that
you need a card that has an UHCI/OHCI AND EHCI compatible chipset.
BTW, did you ever got a faster transfer rate with any other device
(e.g. a card reader)? Even my CF card reader with a SanDisk Ultra II
card easily achieves 6MB/sec when reading and more than 3MB/sec when
writing.
Joe Negron
><jne...@XmindspringX.com>], who wrote in article <QhxLh.11463$PL....@newsread4.news.pas.earthlink.net>:
>> Elapsed Bytes Filesize Xfer rate
>> (secs) (GB) Files avg. (MB) (MB/sec.)
>>
>>SimpleDrive (JFS->JFS) 9,585 61.08 7,915 7.90 6.53
>>SimpleDrive (JFS->JFS) 997 9.67 114 86.86 9.93
>>SimpleDrive (JFS->JFS) 210 1.41 1 1,443.84 6.87
>>Voyager (JFS->HPFS) 190 1.41 1 1,443.84 7.61
>>
>>Can you explain this?
>
>"This" being *what*? Large file giving a smaller transfer speed?
Yes, among other things such as why is the transfer rate to a flash
drive only marginally faster than to a hard disc? Why is the second
test above (copying 114 files with an average filesize of about 87 MB)
significantly faster than test three (copying a single file of 1.41 GB)?
It seems to me that if coping to a flash drive is only marginally faster
than to a hard disc it is likely that I'm getting close to the maximum
for this hardware. The problem is, this maximum doesn't seem to
correspond to the speeds one should get with this interface.
>I do not observe this on my system (copying done using File
>Commander). All I can say is that this should *extremely* sensitive
>to the strategy used by copying program: does it preallocate the space
>for the target? What size of chunks it uses from writing? Does it
>try to interleave reading and writing (multithreaded copy)?
I don't know - I used 4OS2's copy command. I also tested File Commander
and got similar results.
--
------------------------------------------------------------------------
Handy guide to modern science: If it's green or wriggles, it's biology.
If it stinks, it's chemistry. If it doesn't work, it's physics.
Joe Negron
>>>On eCS with the latest USB drivers
>>
>>Which are? I have:
>> 12-21-05 14:13 23,386 usbd.sys
>> 12-21-05 14:14 42,112 usbehcd.sys
>> 12-21-05 14:14 41,984 usbmsd.add
>> 12-21-05 14:14 35,532 usbohcd.sys
>> 7-10-06 4:28 32,326 usbresmg.sys
>
>Yeah, based on the file sizes yours seem to be the same ones.
>
>bldlevel usbd.sys -> 10.162
>bldlevel usbehci.sys -> 10.162
>bldlevel usbuhcd.sys -> 10.162
>bldlevel usbmsd.add -> 10.162
>bldlevel os2dasd.dmd -> 14.104
Yes, they're identical - I checked with bldlevel as well. I also
checked os2dasd.dmd.
>>In case it matters, follows are the USB related entries in CONFIG.SYS:
>>
>> BASEDEV=usbehcd.sys /V
>> BASEDEV=usbohcd.sys /V
>> BASEDEV=usbohcd.sys /V
>> BASEDEV=usbd.sys
>> BASEDEV=usbmsd.add /V /FLOPPIES:0 /REMOVABLES:3
>> DEVICE=C:\OS2\BOOT\usbresmg.sys /V
>
>I have a VIA K8T800 chipset with 8237 southbridge (Athlon64)
>and thus use the UHCI driver.
I have an ASUS A7N8X (nVidia)...
>Some older nVidia chipsets were known to have lousy USB throuput
>but I can't remember which ones.
How would I find out if mine is one of them? (though, it's not really
"old".)
>Well, this really sounds as a hardware issue as also your
>XP makes it not faster.
Yes, this is the only thing that makes sense to me. But, "hardware
issue" could mean different things.
>>>If you get much lower transfer rates, check that in BIOS the
>>>USB 2.0 Highspeed mode is enabled.
>>
>>There exists no option in my BIOS with that precise designation. There
>>is a "USB Controllers" setting, with the following options: "V1.1+V2.0",
>>"V1.1", and "NONE". Of course, I have it set to the first of these.
>
>What chipset is this?
We're getting a little beyond my level of expertise here (e.g., I don't
know what "Northbridge" and "Southbridge" refer to), but the
specifications summary of the motherboard manual states:
Northbridge: NVIDIA nForce2 SPP (Ultra 400)
Southbridge: NVIDIA nForce2 MCP-T (Deluxe Model)/MCP
(note that I have the A7N8X Deluxe)
>Are you sure that it really supports USB 2.0 Highspeed mode
>and not just Fullspeed?
I had no idea there was anything beyond USB 1.1 and USB 2.0! Since
reading your message I've been checking out some USB related info on
wikipedia, www.everythingusb.com, etc.
OK, so there's (from wikipedia):
USB supports three data rates:
* A Low Speed rate of up to 1.5 Mbit/s (187.5 kB/s) that is mostly
used for Human Interface Devices (HID) such as keyboards, mice,
and joysticks.
* A Full Speed rate of up to 12 Mbit/s (1.5 MB/s). Full Speed was
the fastest rate before the USB 2.0 specification and many devices
fall back to Full Speed. Full Speed devices divide the USB
bandwidth between them in a first-come first-served basis and it
is not uncommon to run out of bandwidth with several isochronous
devices. All USB Hubs support Full Speed.
* A Hi-Speed rate of up to 480 Mbit/s (60 MB/s).
But, this doesn't make sense; obviously, I'm getting better than "Low
Speed" and "Full Speed" thus proving I have a USB 2.0 connection.
However, I'm getting substantially less than "Hi-Speed".
IOW, my benchmarks varied between approximately 6.5 to 10 MB/sec.
Therefore, I must have a "Hi-Speed" connection else how would I be
getting 6.5 to 10 MB/sec when "Full Speed" can only do about .75 MB/sec
(unidirectionally). Yet, if I do have a "Hi-Speed" connection, why am I
topping out at only 10 MB/sec when it should approach 30 MB/sec?
>Simply "USB 2.0" says nothing about the
>supported transfer rate. Usually Highspeed capable devices and
>controllers have a separate logo (USB 2.0 Highspeed).
I don't see anything with the Highspeed logo, either on the motherboard
or in the motherboard manual.
>If the board only supports Fullspeed mode (12MBit/sec max.) you can
>try with an add-on USB card that supports USB 2.0 Highspeed mode.
>Even though I haven't done this, the USB driver readme says that
>you need a card that has an UHCI/OHCI AND EHCI compatible chipset.
I came across the SIIG USB 2.0 4-Port Bay Hub:
http://www.siig.com/product.asp?pid=577
which I may get if I can determine that (for whatever reason) my current
hardware is incapable of rates faster than I'm getting.
>BTW, did you ever got a faster transfer rate with any other device
>(e.g. a card reader)? Even my CF card reader with a SanDisk Ultra II
>card easily achieves 6MB/sec when reading and more than 3MB/sec when
>writing.
I'm unsure of what you're asking here. My recent post to Ilya included
the following:
Elapsed Bytes Filesize Xfer rate
(secs) (GB) Files avg. (MB) (MB/sec.)
SimpleDrive (JFS->JFS) 9,585 61.08 7,915 7.90 6.53
SimpleDrive (JFS->JFS) 997 9.67 114 86.86 9.93
SimpleDrive (JFS->JFS) 210 1.41 1 1,443.84 6.87
Voyager (JFS->HPFS) 190 1.41 1 1,443.84 7.61
("Voyager" is a Corsair Voyager 4 GB flash drive formatted as a single
HPFS partition for its entire capacity)
--
------------------------------------------------------------------------
Individuality is either the mark of genius or the reverse. Mediocrity
finds safety in standardization.
--Frederick E. Crane
Ilya Zakharevich
Joe Negron
> >"This" being *what*? Large file giving a smaller transfer speed?
>
> Yes, among other things such as why is the transfer rate to a flash
> drive only marginally faster than to a hard disc?
Why do you expect it to be quickier? A general rule which hold some
time ago was: Flash memory is slow; disks are quick. A particular
year's ratio may fluctuate, but I see no reason to expect flash to
outperform HD.
> It seems to me that if coping to a flash drive is only marginally faster
> than to a hard disc it is likely that I'm getting close to the maximum
> for this hardware. The problem is, this maximum doesn't seem to
> correspond to the speeds one should get with this interface.
Interface is an interface only. If you have a 4-lane highway, you
will not drive it faster than a 3-lane one if you are the only car on
the highway. With quick enough interface, the device's speed becomes
a limiting factor...
Hope this helps,
Ilya
Heiko Nitzsche
> "old".)
I don't think the nForce2 Ultra 400 is one of them.
> We're getting a little beyond my level of expertise here (e.g., I don't
> know what "Northbridge" and "Southbridge" refer to), but the
> specifications summary of the motherboard manual states:
The northbridge mainly maintains the AGP bus and memory transfers
between CPU and RAM while the southbridge is the IO chip for most
of the device ports like USB, LAN, Audio and also PCI cards.
>> If the board only supports Fullspeed mode (12MBit/sec max.) you can
>> try with an add-on USB card that supports USB 2.0 Highspeed mode.
>> Even though I haven't done this, the USB driver readme says that
>> you need a card that has an UHCI/OHCI AND EHCI compatible chipset.
>
> I came across the SIIG USB 2.0 4-Port Bay Hub:
>
> http://www.siig.com/product.asp?pid=577
>
> which I may get if I can determine that (for whatever reason) my current
> hardware is incapable of rates faster than I'm getting.
This is just a hub, not a controller card. It uses the same
on-board USB ports via cable. So it would not make any difference.
This is for instance a controller card:
http://www.siig.com/product.asp?pid=305&catid=18
(the USB 2.0 highspeed logo is below the card picture)
Don't know if it would work. It is just an example.
>> BTW, did you ever got a faster transfer rate with any other device
>> (e.g. a card reader)? Even my CF card reader with a SanDisk Ultra II
>> card easily achieves 6MB/sec when reading and more than 3MB/sec when
>> writing.
>
> I'm unsure of what you're asking here. My recent post to Ilya included
> the following:
>
> Elapsed Bytes Filesize Xfer rate
> (secs) (GB) Files avg. (MB) (MB/sec.)
>
> SimpleDrive (JFS->JFS) 9,585 61.08 7,915 7.90 6.53
> SimpleDrive (JFS->JFS) 997 9.67 114 86.86 9.93
> SimpleDrive (JFS->JFS) 210 1.41 1 1,443.84 6.87
> Voyager (JFS->HPFS) 190 1.41 1 1,443.84 7.61
>
> ("Voyager" is a Corsair Voyager 4 GB flash drive formatted as a single
> HPFS partition for its entire capacity)
Sorry, I have somehow missed this info. This changes the picture a lot.
The transfer rate for the flash drive looks fine. It is realistic.
So a different controller card would not help. Your ports and drivers are fine.
It seems more that the SimpleDrive and your USB controller don't like
each other or the disk drive is simply not faster (which I doubt).
Can't you test with a different hard disk borrowed from a friend?
Maybe try it on XP first (as it also is impacted) if he has no OS/2.
Joe Negron
>[A complimentary Cc of this posting was sent to
>Joe Negron
><jne...@XmindspringX.com>], who wrote in article <T5wMh.130080$_73.8...@newsread2.news.pas.earthlink.net>:
>>>"This" being *what*? Large file giving a smaller transfer speed?
>>
>>Yes, among other things such as why is the transfer rate to a flash
>>drive only marginally faster than to a hard disc?
>
>Why do you expect it to be quickier? A general rule which hold some
>time ago was: Flash memory is slow; disks are quick. A particular
>year's ratio may fluctuate, but I see no reason to expect flash to
>outperform HD.
I'd heard about flash memory being slow, but I thought that meant it was
slow compared to other forms of memory; I assumed flash memory would
easily be fast enough to be faster than anything mechanical.
>>It seems to me that if coping to a flash drive is only marginally faster
>>than to a hard disc it is likely that I'm getting close to the maximum
>>for this hardware. The problem is, this maximum doesn't seem to
>>correspond to the speeds one should get with this interface.
>Interface is an interface only. If you have a 4-lane highway, you
>will not drive it faster than a 3-lane one if you are the only car on
>the highway. With quick enough interface, the device's speed becomes
>a limiting factor...
OK, you're saying the SimpleDrive is the limiting factor. As I've
previously mentioned, the documentation states that it could get "up to
480 Mb/sec" for USB 2.0. So, either the documentation is incorrect, or
there's something wrong with my setup.
At this point it seems I need to get some feedback from others who have
this drive.
--
------------------------------------------------------------------------
Difficulties strengthen the mind, as labor does the body.
--Seneca
Joe Negron
>>>If the board only supports Fullspeed mode (12MBit/sec max.) you can
>>>try with an add-on USB card that supports USB 2.0 Highspeed mode.
>>>Even though I haven't done this, the USB driver readme says that
>>>you need a card that has an UHCI/OHCI AND EHCI compatible chipset.
>>
>>I came across the SIIG USB 2.0 4-Port Bay Hub:
>>
>> http://www.siig.com/product.asp?pid=577
>>
>>which I may get if I can determine that (for whatever reason) my current
>>hardware is incapable of rates faster than I'm getting.
>This is just a hub, not a controller card. It uses the same
>on-board USB ports via cable. So it would not make any difference.
>
>This is for instance a controller card:
>http://www.siig.com/product.asp?pid=305&catid=18
>(the USB 2.0 highspeed logo is below the card picture)
>Don't know if it would work. It is just an example.
So, if I were to get a controller card, I'd have to disable the onboard
USB ports? If so, would that simply involve setting the "USB
Controllers" BIOS option to "NONE" (instead of "V1.1" or "V1.1+V2.0")?
>>>BTW, did you ever got a faster transfer rate with any other device
>>>(e.g. a card reader)? Even my CF card reader with a SanDisk Ultra II
>>>card easily achieves 6MB/sec when reading and more than 3MB/sec when
>>>writing.
>>I'm unsure of what you're asking here. My recent post to Ilya included
>>the following:
>>
>> Elapsed Bytes Filesize Xfer rate
>> (secs) (GB) Files avg. (MB) (MB/sec.)
>>
>>SimpleDrive (JFS->JFS) 9,585 61.08 7,915 7.90 6.53
>>SimpleDrive (JFS->JFS) 997 9.67 114 86.86 9.93
>>SimpleDrive (JFS->JFS) 210 1.41 1 1,443.84 6.87
>>Voyager (JFS->HPFS) 190 1.41 1 1,443.84 7.61
>>
>>("Voyager" is a Corsair Voyager 4 GB flash drive formatted as a single
>>HPFS partition for its entire capacity)
>Sorry, I have somehow missed this info. This changes the picture a lot.
>
>The transfer rate for the flash drive looks fine. It is realistic.
>So a different controller card would not help. Your ports and drivers are fine.
>
>It seems more that the SimpleDrive and your USB controller don't like
>each other or the disk drive is simply not faster (which I doubt).
>
>Can't you test with a different hard disk borrowed from a friend?
>Maybe try it on XP first (as it also is impacted) if he has no OS/2.
Unfortunately, that's not an option. I'm going to try to find some
feedback, somewhere online, from others with this drive. I still have
time to return it if I determine the problem to be simply a slow drive.
I'd rather pay more (though not a whole lot more) for a faster drive; if
I'm fortunate a faster drive will be fast enough so it can be formatted
FAT32 and still give me the appoximate speed I'm now getting.
--
------------------------------------------------------------------------
Absence diminishes little passions and increases great ones, as wind
extinguishes candles and fans a fire.
--François de La Rochefoucauld
Ilya Zakharevich
Joe Negron
> OK, you're saying the SimpleDrive is the limiting factor. As I've
> previously mentioned, the documentation states that it could get "up to
> 480 Mb/sec" for USB 2.0. So, either the documentation is incorrect, or
> there's something wrong with my setup.
I see no reason to make this conclusion. Does the documentation state
that it can reach 480 Mb/sec for sustained transfer?
(Cheap disk drives of today can easily reach 70MB/sec sustained.)
Hope this helps,
Ilya
Peter Brown
Joe Negron wrote:
> On 2007-03-22, Heiko Nitzsche <hn-expire...@arcor.de> wrote:
>>>> If the board only supports Fullspeed mode (12MBit/sec max.) you can
>>>> try with an add-on USB card that supports USB 2.0 Highspeed mode.
>>>> Even though I haven't done this, the USB driver readme says that
>>>> you need a card that has an UHCI/OHCI AND EHCI compatible chipset.
>>> I came across the SIIG USB 2.0 4-Port Bay Hub:
>>>
>>> http://www.siig.com/product.asp?pid=577
>>>
>>> which I may get if I can determine that (for whatever reason) my current
>>> hardware is incapable of rates faster than I'm getting.
>> This is just a hub, not a controller card. It uses the same
>> on-board USB ports via cable. So it would not make any difference.
>>
>> This is for instance a controller card:
>> http://www.siig.com/product.asp?pid=305&catid=18
>> (the USB 2.0 highspeed logo is below the card picture)
>> Don't know if it would work. It is just an example.
>
> So, if I were to get a controller card, I'd have to disable the onboard
> USB ports? If so, would that simply involve setting the "USB
> Controllers" BIOS option to "NONE" (instead of "V1.1" or "V1.1+V2.0")?
>
I suggest that you test that.
I have only installed 1 add-in USB2 controller card to a system with
existing USB1 controllers. Disabling the mainboard controllers in the
BIOS resulted in the add-in card not working as well. So, the system
with the add-in controller card had 1 ehci, 3 ohci and 2 uhci
controllers all working.
I think you need to Enable a BIOS setting called something like "Update
ESCD" when booting after adding the controller card otherwise there is a
chance that it may not be recognised by OS/2.
Regards
Pete