gpsd-users
[Top][All Lists]
Advanced

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

Re: [gpsd-users] How to get the best performance out of gpsd?


From: Gary E. Miller
Subject: Re: [gpsd-users] How to get the best performance out of gpsd?
Date: Fri, 16 Aug 2019 10:57:38 -0700

Yo pisymbol!

On Fri, 16 Aug 2019 09:10:13 -0400
"pisymbol ." <address@hidden> wrote:

> On Thu, Aug 15, 2019 at 4:18 PM Gary E. Miller <address@hidden> wrote:
> 
> > Yo pisymbol!
> >
> > On Thu, 15 Aug 2019 09:10:48 -0400
> > "pisymbol ." <address@hidden> wrote:
> >  
> > > Or I think what might be more lightly than all of the above is the
> > > sample() call is happening at irregular rates.  
> >
> > Duh.  Of course it is irregular.  No GNSS receiver outputs
> > consistent NMEA streams.  gpsd just passes on what it gets, when it
> > gets it. 
> 
> 
> Note: I sent you code in my last email, not sure why you didn't see
> get it?

I did not see complete, working, stand alone code.  If you sent that,
please try again.

> The question is how irregular?

Depends on the GPS.  TPV calculations are doen with iterative
calculations from the raw sat observation data, so they take vaying
amounts of time.  Then the NMEA is vayring length, so also takes varying
lengths of time to transmit.  You can see this by turning on the timeing
flag when you you start your WATCH.

Or use the -t option to gpspipe.


> But more importantly, looking at the
> gps/client.py and gps/gps.py files, it seems that the next iterator
> devolves into a sock.recv() (L115) of 8k.

Yes.  Did you read the man page for that?  "man libgps"?

> Question: Why isn't it using
> select I/O itself like gpspipe?

Many ways to skin a cat.  The examples intentionally use different methods
to show the differences.

> Does that turn out to be slower?

Always fast enough.  Have you tried using cgps and gpspipe to see how
that work in your setup?

> I
> suspect what is happening is occasionally on the Trimble side, it's
> takes a long time for it to emit NMEA strings for whatever reason.

I doubt it.  But cgps and gpspipe will show you if it does in fact
happen.

> I'm still working out some metrics though.

No need, use the ones we give you.  There are lot's of them.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
        address@hidden  Tel:+1 541 382 8588

            Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

Attachment: pgp_AOFXtRUVa.pgp
Description: OpenPGP digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]