Experimental Physics and
Industrial Control System

Jeong Han Lee via Tech-talk <[email protected]> · Wed, 3 Jul 2019 14:57:22 -0400

Hi Andrew,

 I am really interested to resume old codes to build a standard

procedure for latest release. Could you share any technical

documents/source codes which were used for this test.

  I am reading very old (not really) 1993 technical document now.

  Best,
  Han

On 7/1/19 2:45 PM, Johnson, Andrew N. via Tech-talk wrote:

Hi Han,

On 6/30/19 12:42 PM, Jeong Han Lee via Tech-talk wrote:

Is there a standard EPICS CA/PA or V3/V4/V7 performance test scenario

or tools, or even results?

There have been various performance measurements made over the years,

and we even collected some results about individual VME CPU boards, but

those results aren't terribly meaningful nowadays (the last result

recorded was measured in 2007) and the code we used will probably need

work to get it to build on recent EPICS releases. This thread

<https://epics.anl.gov/tech-talk/2013/msg00646.php> has some information

about that application, the measurements are in the performApp/TIMING

text file.

In other form of question, "do we have the standard procedure to see

deployed IOC limitation?"

No. I guess most people see whether it's fast enough for their

application and restructure or adjust it until it is.

In other form of question, "do we have a benchmark tool for any IOC or

application?"

Base builds a program called catime which measures the performance of a

CA client/server connection (there are some instructions on

<https://epics.anl.gov/base/R7-0/2-docs/CAref.html#catime> how to run it

in the CA reference documentation).

Other tests have been performed by various people over the years and

described in tech-talk messages or talks at collaboration meetings, but

we don't have a standard as such.

HTH,

- Andrew

--
Complexity comes for free, Simplicity you have to work for.

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System