Hi Mark and other friends,
Sorry to cause all the noise. I think it is working now.
This is one of my own records, it has some floats and some arrays fields. For the
array which I would like to monitor, I created 2 buffers and let cvt_dbaddr assign one
for user to access while the other is being filled up, then switch. Then I had another
complication: I also wanted to monitor a float field. so I was doing:
db_post_event ( for the float)
db_post_event ( for the array whose address toggles between 2 candidates)
Anyway, when I removed the monitoring for the float, and implemented a single
buffer for the array (now I have to memcpy), it seems to be working. I will try
some combinations to see what exactly was causing all the trouble.
Thank you all very much! Best regards,
Dehong
----- Original Message -----
From: Mark Rivers <[email protected]>
Date: Monday, August 11, 2008 1:45 pm
Subject: RE: Loosing half the monitor callbacks
To: Dehong Zhang <[email protected]>, Tim Mooney <[email protected]>
Cc: [email protected]
> You need to tell us:
>
> - What records are you using? Are these standard EPICS records, or
> records you have written.
>
> - What device support are you using? Did you write it?
>
> I don't believe there is any problem with posting monitors on waveforms
> at >10Hz, I do it frequently and I don't think I am having any problem.
>
> I suspect there is a problem in your record support or device support.
>
> Can you send the code where the calls to db_post_events is done?
>
> Mark
>
>
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of Dehong Zhang
> Sent: Monday, August 11, 2008 1:34 PM
> To: Tim Mooney
> Cc: [email protected]
> Subject: Re: Loosing half the monitor callbacks
>
> Hi Tim (and Jeff),
>
> There seems to be a little imperfection somewhere.
>
> In my record, I try to "monitor" 2 fields: a float and an array of
> floats
> (288 elements). What I want is to get updates at about 5Hz. It does
> not work -- missing every the other update but get those lucky ones
> twice -- still getting 5 Hz but half the information.
>
> I tried to slow down the system to about one update every 3 seconds.
> It still behaves the same way. However when I commented out the
> db_post_event for the array, the monitoring for the single float works
> correctly. The problem seems to be related to the array.
>
> In the Tech-Talk DB, Jeff once replied to someone about buffering
> in Epics (the guy had very similiar problem). Am I hit by the same
> thing?
> It is hard to imagine that at 0.3 Hz I would run into buffer-full or
> out-
> of-memory problems. I would imagine that at a slow enough rate,
> it should work with whatever buffering or flowControl.
>
> Is there a way to re-config the IOC so that it can 'monitor' small
> arrays
> at <= 10 Hz? My processor has 500 MB memory.
>
> Thank you very much. Best regards,
> Dehong
>
>
>
>
> ----- Original Message -----
> From: Tim Mooney <[email protected]>
> Date: Saturday, August 9, 2008 11:17 am
> Subject: Re: Loosing half the monitor callbacks
> To: Dehong Zhang <[email protected]>
> Cc: [email protected], [email protected]
>
>
> > Dehong,
> > I think this shows that the problem is not in channel access, but in
> the
> > database or lower. You're not losing monitor callbacks, the record
> is
> > posting the same value twice, with the same time stamp. You can get
> > another clue to what this means by looking at the record-support code,
> > and finding calls to recGblGetTimeStamp() and calls to
> db_post_events()
> > for the CTME field.
> >
> > Dehong Zhang wrote:
> > > Hi Friends Kay, Mark and Mohan,
> > >
> > > Thank you very much for the advices.
> > >
> > > I just tried camonitor. It is doing the same thing: throw away
> > every the other
> > > change, but callback two times for the other ones. The printout
> is
> > like:
> > >
> > > fastTimeTest.CTME 2008-08-08 16:46:02.609123
> > 6.88202e+06
> > > fastTimeTest.CTME 2008-08-08 16:46:02.859123
> > 7.38478e+06
> > > fastTimeTest.CTME 2008-08-08 16:46:02.859123
> > 7.38478e+06
> > > fastTimeTest.CTME 2008-08-08 16:46:03.259123
> > 8.01506e+06
> > > fastTimeTest.CTME 2008-08-08 16:46:03.259123
> > 8.01506e+06
> > > fastTimeTest.CTME 2008-08-08 16:46:03.659123 137329
> > > fastTimeTest.CTME 2008-08-08 16:46:03.659123 137329
> > > fastTimeTest.CTME 2008-08-08 16:46:04.059123 817524
> > > fastTimeTest.CTME 2008-08-08 16:46:04.059123 817524
> > >
> > > Does the record update time stamps tell anything? Am I running
> too
> > fast?
> > >
> > > Thank you again. Best regards,
> > > Dehong
> > >
> > >
> > >
> > >
> > >
> > > ----- Original Message -----
> > > From: Mohan Ramanathan <[email protected]>
> > > Date: Friday, August 8, 2008 4:43 pm
> > > Subject: RE: Loosing half the monitor callbacks
> > > To: Dehong Zhang <[email protected]>, [email protected]
> > >
> > >
> > >> Hi
> > >> Try to set the TPRO field in the IOC to see that it is indeed
> > >> processing every time
> > >>
> > >> Mohan
> > >>
> > >> ----------------------------------------------------------------
> > >> Mohan Ramanathan ([email protected])
> > >> Senior Beamline Engineer, Advanced Photon Source
> > >> Argonne National Laboratory, Argonne, IL 60439
> > >> (630)-252-3773/ (630)-822-5625(Cell)
> > >> ----------------------------------------------------------------
> > >>
> > >> [Message delivered by NotifyLink]
> > >>
> > >> ----------Original Message----------
> > >>
> > >> From: Dehong Zhang <[email protected]>
> > >> Sent: Fri, August 08, 2008 4:30 PM
> > >> To: [email protected]
> > >> Subject: Loosing half the monitor callbacks
> > >>
> > >>
> > >> Hi,
> > >>
> > >> I would like to monitor a PV at 5 Hz. Somehow I do get 5 callbacks
>
> >
> > >> every seconds
> > >> but only get the changes every the other time, like getting changes
>
> >
> > >> 1st, 3rd, 5th ...
> > >> two times, but never get changes 2nd, 4th, ...
> > >>
> > >> Could you please enlighten me?
> > >>
> > >> Thanks and best regards,
> > >> Dehong
> > >>
> >
> > --
> > Tim Mooney ([email protected]) (630)252-5417
> > Beamline Controls & Data Acquisition Group
> > Advanced Photon Source, Argonne National Lab.
- References:
- Re: Loosing half the monitor callbacks Dehong Zhang
- RE: Loosing half the monitor callbacks Mark Rivers
- Navigate by Date:
- Prev:
RE: Loosing half the monitor callbacks Mark Rivers
- Next:
Scan Time Definitions David Dudley
- Index:
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
<2008>
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
- Navigate by Thread:
- Prev:
RE: Loosing half the monitor callbacks Mark Rivers
- Next:
Scan Time Definitions David Dudley
- Index:
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
<2008>
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
|