Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Wed, 11 Jan 2012 11:30:00 -0600

Hi Anton,

On 2012-01-11 Anton Derbenev wrote:
> When I add a callback for a PV object with array
>  data within, I become "subscribed" to the entire array. The problem may be
>  that, as described below, the PV image data is always seen as large as (in
>  my case) 1228800 elements array, while the actual region of interest size
>  can be much smaller (all irrelevant elements of the array simply being set
>  to 0). So, is it possible to create the PV object and the callback for it
>  that will only monitor the specified part of itself (like, I assume,
>  camonitor -# does)?

Assuming that you're using EPICS Base version 3.14.12.1 or later the 
underlying CA library and IOC server do support variable length arrays, but I 
don't know if or how pyepics handles them.  When calling the underlying CA 
library directly from C code if you request an array of zero elements in a 
get-callback or monitor request, you'll get sent the number of elements that 
currently appear in the IOC's array field with no zero padding added.  This 
approach is what the camonitor program uses, and if pyepics supports it you'll 
want to use this feature for maximum throughput of array data.

Note that cainfo reports the size of the storage allocated for the array on 
the IOC, which is the maximum size the array can be.  It can't tell you how 
much data the array currently holds.

HTH,

- Andrew
-- 
Optimization is the process of taking something that works and
replacing it with something that almost works, but costs less.
-- Roger Needham

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System