Experimental Physics and
Industrial Control System

"Mark Rivers" <[email protected]> · Tue, 31 May 2011 10:08:09 -0500

Hi Dirk,

I think there are 3 problems:

1) I think lailongwei is running caget 4 times to get 4 PVs.  Perhaps he
is not aware that he can get 4 PVs running caget only once:

corvette:> /usr/local/epics/base-3.14.12.1/bin/linux-x86/caget 13IDA:m1
13IDA:m2 13IDA:m3 13IDA:m4
13IDA:m1                       -0.25007
13IDA:m2                       1.19993
13IDA:m3                       -0.649928
13IDA:m4                       0.49989

2) Your point that even those values are not truly sampled at the same
time.

3) The problem he is probably really complaining about, which is that
caget it taking 10 seconds to complete, whereas when things are working
well it should take less than 50ms to complete.  This means that there
is something seriously wrong with caget, and he is not the first one to
report this.  I have seen it take up to 5 seconds to complete from some
hosts to some IOCs, as has Jay Steele.

Mark

-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Dirk Zimoch
Sent: Tuesday, May 31, 2011 7:14 AM
To: lailongwei
Cc: tech-talk
Subject: Re: caget command holds for more than 10 seconds

lailongwei wrote:
> hi,
>   when I use caget A, though the data has been gotten, it holds for 
> about ten sends untill the command exit.
> This cause problem when I need to caget four PVs, whose sampling at
the 
> same time is very important. How can I solve it.
> Thanks!
>  
> 2011-05-31
>
------------------------------------------------------------------------
> lailongwei

One could argue that strictly speaking sampling over the network is 
"wrong". TCP and thus CA does not support any real-time behavior. So the

best way to solve it is to sample all four PVs on the same real-time 
IOC. This probably does not help when the PVs are not physically 
connected to the same IOC. For the future SwissFEL project, we (meaning 
Babak) took great effort to implement this kind of synchronous data 
acquisition across multiple IOCs. Our implementation strongly depends on

our timing system to synchronize PVs on multiple IOCs. This is nothing 
you can quickly do with a 'caget a; cage b; caget c; caget d;'

Dirk

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System