Experimental Physics and
Industrial Control System

Susanna Jacobson <[email protected]> · Fri, 14 Nov 2003 11:44:31 -0800

Steven Hartman wrote:
>
>  - Is there anything in base which may break or otherwise behave
> differently if the VxWorks system clock is increased to be faster than 60
> Hz? Looking at dbScan.c, it looks like the periodic scan rate will do the
> right thing (so long as sysClkRateSet() is called before initPeriodic(),
> so before iocInit()). drvTS.c also looks OK. Anything else in base that
> may be a problem? (I will of course check device code, sub records, etc
> which are used in the app which are not in base.)

The ALS control system includes 15 MCP750 (compact PCI Power PC, 3 235MHz 
and 12 367MHz) IOCs with a system clock rate of 100Hz that have been running 
with no problems for > 2 yrs.

>  - Does anyone have experience on a maximum reasonable VxWorks clock rate
> with an EPICS application? Same caveats as above, but a rough guideline
> would be useful.

We tried running the 12 367MHz CPUs with a clock rate of 1000Hz, but then 
we saw network operations running out of memory buffers, causing channel 
access connections to be dropped.  No testing at intermediate clock rates 
was done.  So my guess is that the maximum reasonable rate is somewhere 
between 100Hz and 1000Hz.

	Susanna
-- 
=====================================================================
Susanna Jacobson                MS 10-110
[email protected]              Lawrence Berkeley National Laboratory
VOICE:  (510) 486-7801          1 Cyclotron Road
FAX:    (510) 486-7374          Berkeley, CA 94720
=====================================================================

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System