Experimental Physics and
Industrial Control System

Ralph Lange <[email protected]> · Tue, 23 Jan 2018 14:12:34 +0100

On Tue, Jan 23, 2018 at 10:13 AM, Anton Derbenev <[email protected]> wrote:
>
> It worth mentioning that even though my simulation IOCs are relatively heavy
> on amount of computations they do (lots of chained waveform calculations and
> aSubs), they work very well even on x100 rate without any apparent issues. I
> don't know what is the lowest scan rate achievable with the described method
> though. I guess minimal reasonable value depends on an application.

Technically, the lowest reasonable SCAN rate is limited by the
resolution of the system timer.
(Historically, the VxWorks clock was ticking at 60Hz, so 10Hz was a
reasonable choice.)

If you are able to run an IOC application using a fast rate obviously
depends on the application and the CPU power of the machine the IOC is
running on.
Note for multi-core machines: while running multiple callback threads
(that do 'I/O Intr' processing) in parallel has been available for a
while, there is still one periodic scan thread per SCAN period. Adding
support for parallel SCAN threads is conceptionally possible, but
needs work to be done (or funded).

Cheers,
~Ralph

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System