EPICS Controls Argonne National Laboratory

Experimental Physics and
Industrial Control System

<19941995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024  Index <19941995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
<== Date ==> <== Thread ==>

Subject: Re: IOC CPU recommendations
From: [email protected]
Date: Wed, 01 Jun 94 15:14:54 -0500
>We need to replace IBM PC clones (with Bitbus interfaces) with VME
>systems for the IOCs.

Interesting to see you are using bitbus.  As of late, we have had a
giant problem with what SEEMS to be due to a protocol error recovery
that the bitbus spec has defined.

We end up in a situation where we send message N and get the response
from message N-1 to a given slave node.  And this is at the bitbus
interface level... not a silly driver error.  We are using both the
xycom and PEP modular bitbus controllers.  And it is reproducable on
both.  It is driving me nuts.  I'd like to talk to you (or anyone else
using bitbus) about how your driver's API insures your responses match
up with the original messages.


On with the real business...

>The 162 seems to be limited to 4MB RAM...

Yes, on the CPU this has been true so far.  I know nothing of future
plans, but I suspect larger memories will be soon available.

As an alternative, you can easily add memory using the IP boards.

At the APS, our databases have been running on 8MB 167s and we have run
into some memory problems when a number of CA clients are attached. 
These databases have in excess of 1000 records in them.  Your milage
will certainly vary.

>Fred Carter @ DUKE mentioned to my boss (Coles Sibley) that there
>have been problems with the 167.  Is there still reason to be
>concerned about this?

The only problems I have herd of with the 167s is that there was a
recall a while ago where they had some problem with the TAS operation
across the VME backplane.

All other problems I am aware of have been due to the choice of REALLY
STUPID defaults in the vxWorks BSP code.  Much of those problems have
been worked out.  I am sure that more will come up as time goes on, but
that is going to happen no mater what board you choose.

>Does the 167 give us anything else that the 162 doesn't? (We know
>the 162 has Industry Pack capabilities, but we don't feel this
>will be that useful to us)

The only serious differences between the two are the memory
interfaces... 162 is onboard only, and 167 is add-on boards.  The 162
has IP, and 167 does not.  And the serial interface on the 162 has 2
ports, and the 167 has 4.  (In fact, I would bet that the serial stuff
is the only reason that the BSPs are not the same between the two
boards.

You should also look into the 177s.  They are 68060 based and should
run faster than the 167... like 30%, says Moto.  And they are about
$500 more expensive.  wind rivers is working on the BSP at this moment.
The 177 has all the exact same stuff as the 167, but since the 060 does
not have a MSP register (and a few other possible problems) the regular
040 BSP aint gonna cut it.  I do not know the status of this new BSP
yet.  But I suspect that for the non-IP user, the 177 would be the best
bet.

You might also want to inquire about these boards on the
comp.os.vxworks news group.  There are several users out there that
have been talking about the 167 and 162 boards.

--John Winans

P.S.  If you are going to be at the EPICS user meeting out here in two
weeks, find me and I will show you what I have on the 162 and 167
boards.

Navigate by Date:
Prev: IOC CPU recommendations Tony Carter, MIT Bates Linac
Next: Re: IOC CPU recommendations Tim Mooney
Index: <19941995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
Navigate by Thread:
Prev: IOC CPU recommendations Tony Carter, MIT Bates Linac
Next: Re: IOC CPU recommendations Tim Mooney
Index: <19941995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
ANJ, 10 Aug 2010 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· Search · EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·