Re: DateBk5 Categories Support / Speed of Synch



On Sun, 2003-06-15 at 20:04, David A. Desrosiers wrote:
> 	Setting the speed to 115200 under USB has no effect, unless you're
> using a "real" RS-232 serial port. If you want to try some additional
> speedups, try using a port designation of 'usb:/dev/ttyUSB1' (no
> double-slash, just one), and see if that helps. You have to be running a
> recent pilot-link to enable it, and of course, a recent gnome-pilot as well.

David,

Thanks for your reply.

As you may have noted in my reply earlier his afternoon, I have upgraded
to the most recent versions of both gnome-pilot/conduits (2.0.9) and
pilot-link (0.11.7).

A question, I understand the logic of the protocol specification you
have provided above, however, I cannot enter this into the gnome-pilot
GUI as it will not permit ':' characters in the field. Is there another
way to enter this or is there another place that I should be specifying
the 'usb:/dev/ttyUSB1' port?

As you may have noted I am presently using /dev/usb/ttyUSB1.


> 	On the right hardware, it will give you a 50% speedup in sync time.
> Basically that switch will remove the unnecesary layer of the DLP protocol
> stack, resulting in faster transfers. Some people report success, some
> people report failures with it, YMMV.
> 
> > Is the synch time a reflection of slower net USB speeds under RH 9 or is
> > there some feature in Outlook (ie. a changed record flag) that enables the
> > synching to be faster than with Evolution? If so, then again, this may
> > raise the question of this being a conduit and/or Evolution issue.
> 
> 	If you are _only_ sync'ing to gnome-pilot, multiple times in a row,
> it should be fast. If you sync to Windows, then gnome-pilot, it will almost
> always be slow, as it should be, because the record pointers and LastSyncPC
> and SyncDate have all changed.


I am synching both to Windows and to Evolution, however, even when
synching twice or three times in a row to Evolution, without Windows in
that sequence, there is no appreciable speedup.  It takes several
minutes for each synch, whereas under Windows if no data has changed, I
can synch and update Avantgo and Documents to Go in about a minute.

Unless, I am missing something, based upon your statement in the last
paragraph above, this suggests a USB bandwidth bottleneck under Linux.
If we can figure out how to make the change in protocol above, I'll
report back my findings regarding performance.

Any further guidance on this would be appreciated.

Regards,

Marc





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]