Re: USB Port speeds

Hi,

> The format
> option of -spare=none did help to at least double the performance as you
> thought.  I was able add the parameter to the command in K3B settings as
> well.

At least with previously used BD-RE media it will not help to
have this option with growisofs. It is intended to prevent
auto-formatting of blank BD-R media. There it is beneficial indeed.

BD-RE must be stripped of their Spare Area by dvd+rw-format.


> the search
> button in K3B would not offer any other choices than what was provided.

They ignore my programs since years. (Shrug)


> So the performance doubled as you expected but still nowhere near what the
> DVD performance was able to get. (4500 KB/s vs. 14000 KB/s)

2x BD speed should be about 9000 KB/s.
4500 still looks like checkreading.

What speed reports do you get from

  dd if=/dev/zero bs=1M count=1000 | \
   xorriso -as cdrecord -v dev=/dev/sr0 -

and from

  dd if=/dev/zero bs=1M count=1000 | \
   xorriso -as cdrecord -v dev=/dev/sr0 stream_recording=on -


> Would I still have the same problem if directly attached to a SATA port?  It
> sounds like I would.

If you can test it then we would know whether the USB driver
is involved.

With my own USB attached drives i have problems when speed
gets to 12x DVD which is about 16 MB/s. On a slightly more
modern machine, the same USB boxes run up to 20x DVD speed.

On my olde machine, the block device driver can read with 
about 20 MB/s whereas my userspace MMC driver gets only
16 MB/s. There seem to be timing and priority problems.
A SATA attached DVD burner at the old machine does 20x
with my userspace driver.

But that all happens far above 10 MB/s.


> At this point I am not sure as to what is the best direction, look for
> another drive or different media.  Any thoughts from your past performance /
> experiences?
> Would drive firmware update help this issue (if available)?

I don't think it is a matter of the BD-RE media if they are
really formatted to 0 Spare area. Bad media would cause write
errors or read errors but should not be slow if Defect Management
is sucessfully disabled.

To verify that state, run
  xorriso -outdev /dev/sr0 -list_formats
It should report before a list of format proposals:
  Format status: formatted, with 23866.0 MiB
  BD Spare Area: 0 blocks consumed, 0 blocks available


> I also have 6x BD-R media but since they are not erasable I prefer not to
> just waste them on silly tests to determine what performance I can get.

There is the "CLOSE SESSION" bug of growisofs with BD-R.
You may want to try my fix proposal for growisofs_mmc.cpp as of
  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=713016
before or after you got hit by that bug.

When it seems necessary to waste BD-R, then you may leave them
appendable (by xorriso -as cdrecord option -multi) and make more
than one experiment with the same medium.
About 22 sessions of 1 GB should fit. Probably one could reduce
session size to a few hundred MB and still get meaningful
measurements.

But for now, 2x BD-RE is a good goal.


Have a nice day :)

Thomas


-- 
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
Archive: http://lists.debian.org/[email protected]