Experimental Physics and
Industrial Control System

"Jeff Hill" <[email protected]> · Tue, 26 Jul 2011 09:15:55 -0600

> The vxWorks "spy" (and "spyStop") commands are also very useful for seeing
> what tasks are using the CPU.

BTW, for this type of thing I frequently use spyClkStart, spyReport, and
spyClkStop to avoid spy's output occurring asynchronously when I am typing
other shell commands.

And yes, the ca server definitely runs at priority below the sequencer, and
also below most of the other components in the IOC.

Jeff
______________________________________________________
Jeffrey O. Hill           Email        [email protected]
LANL MS H820              Voice        505 665 1831
Los Alamos NM 87545 USA   FAX          505 665 5107

Message content: TSPA

With sufficient thrust, pigs fly just fine. However, this is
not necessarily a good idea. It is hard to be sure where they
are going to land, and it could be dangerous sitting under them
as they fly overhead. -- RFC 1925

> -----Original Message-----
> From: [email protected] [mailto:[email protected]]
> On Behalf Of Mark Rivers
> Sent: Tuesday, July 26, 2011 5:35 AM
> To: Dalesio, Leo; Jay Steele; [email protected]
> Subject: RE: question regarding VxStats-1-7-2h (doesn't seem to measure
> CPUloadproperly)
> 
> The vxWorks "spy" (and "spyStop") commands are also very useful for seeing
> what tasks are using the CPU.
> 
> Mark
> 
> 
> ________________________________
> 
> From: [email protected] on behalf of Dalesio, Leo
> Sent: Tue 7/26/2011 6:04 AM
> To: Jay Steele; [email protected]
> Subject: RE: question regarding VxStats-1-7-2h (doesn't seem to measure
> CPUloadproperly)
> 
> 
> A state transition program that transitions to itself with no delays uses
> all CPU time that is at the SNL and lower priorities.
> I do not recall if the Channel Access Client runs at a lower priority than
> the state notation engine.
> I think that the vxStats gotten from vxWorks are computed in the idle
> task. That would mean that it would never run.
> The better tool for finding this problem is to run the SNL diagnostic,
> seqShow "state set name".
> The current state and the previous state would show the two states (or one
> if it is to itself), in the infinite loop. And the time in this state will
> always be zero - or very close to zero.

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System