[comp.realtime] Request for VMEexec informations - long

sc1@lln-cs.UUCP (Alain H. Ninane) (07/23/90)

X-Summary : 1. Do you have any personnal experience with VMEexec ?
	    2. Do you know any VMEexec software ftp server (or
		any other information medium) ?

Hi Folks,

I want to develop a data acquisition system for small-to-medium
sized nuclear physics experiments (1-100 detectors, 200 kb/s max data
throughput). The data will be read by VME processors from CAMAC, VXI
and VME modules and directly written on tape (ExaByte). In addition,
a sample of the data will go, through Ethernet, to UNIX/Ultrix
workstations (both SUN/DEC) for monitoring. One workstation will be
the master controller of the experiment and will send/receive requests
to/from the VME processors. A major requirement is that I want a VME crate
dedicated to the data acquisition task; that means that *I do not want
to see a UNIX (or others high level OS, with disks, screens, and ...)*
in this crate.

From the previous description, most of you will probably declare
that VxWorks is the solution I'm looking for ... but there are
problems, their support of DEC platforms and their licence policy :
VxWorks is too expensive for our budget.

Therefore, we have decided to look to another RTOS : VMEexec. We plan
to develop the software with their combined UNIX System V/68  - VMEexec
development system and then, when this will be finish, to move the VME
processors to the data acquisition crate. Therefore the targets
processors (MVME147 family) should be able to access the Ethernet
by themselves, without the UNIX SV/68 system. Is there any VMEexec
software support for that ? Will it be possible to boot the system
from the a SUN or DEC WorkStation on the net ?

Is there anybody having some personnal experiences with VMEexec ?
Do you know source of VMEexec software ? Do you know if there are
any support for non-Motorola CPU's available for VMEexec ?

Please send your answers to : Ninane@fynu.ucl.ac.be.
I will summarize.

Thanks a lot,

	Alain H. Ninane		E-mail : Ninane@fynu.ucl.ac.be

edski@oz.UUCP (Ed Skinner) (07/23/90)

In article <990@lln-cs.UUCP> Ninane@fynu.ucl.ac.be (Alain H. Ninane) writes:
>X-Summary : 1. Do you have any personnal experience with VMEexec ?

     I am the course developer and one of the instructors for Motorola's
courses on system development with VMEexec.  I also participated as a
development engineer during the early stages of VMEexec definition and
product selection, refinement, and tool creation.

>	    2. Do you know any VMEexec software ftp server (or
>		any other information medium) ?

     The current release (version 1.1) of VMEexec permits target nodes
to manipulate and use global (EXPORTed) objects, such as message queues,
semaphores, memory partitions, and tasks, on any other node.  With the
addition of the "NETlib" software (a separate package from Motorola),
nodes may be interconnected via Ethernet as well as by the VMEbus back-
plane (the "utility" of a global memory partition in an Ethernet config-
uration is, of course, questionable).
     A subsequent release (version 2.0), scheduled for the near future
(I am dialed-in at the moment as we are presenting the class this week
and I don't have all of my notebooks available--sorry) will include a
socket-library for application programs as well as other networking
capabilities.

> [text deleted...]
>
>                                          Therefore the targets
>processors (MVME147 family) should be able to access the Ethernet
>by themselves, without the UNIX SV/68 system. Is there any VMEexec
>software support for that ? Will it be possible to boot the system
>from the a SUN or DEC WorkStation on the net ?

     The VMEexec software support you describe will be available in the
version 2.0 release.  Because it is scheduled "soon," it is best if you
contact our sales folks in Europe to obtain delivery dates.  Note that
you will require VMEexec 2.0 and NETlib 2.0 (which may be called SOCKlib
by the time it actually goes out).
     VMEexec nodes may be booted from either of two places;
1) a SYSTEM V/68 (Unix) host system, or
2) ROM.
     There is an intermediate, mixed case where a node boots from ROM
and then downloads, via Ethernet, a replacement kernel.  However, this
is not what you are specifically asking about.
     For the described configuration, the ROM'd alternative would appear
to be the best.

>Is there anybody having some personnal experiences with VMEexec ?

     No doubt Alain would appreciate hearing from some real customers
rather than from my admittedly biased perspective.

>Do you know source of VMEexec software ? Do you know if there are
>any support for non-Motorola CPU's available for VMEexec ?

     The "vmexgen" utility is used to define, configure, and generate
kernels and network routing information.  VMEexec comes with definitions
for various Motorola CPU boards (about a dozen, I think).  It is normally
a simple matter to use vmexgen to "Copy" one of the Motorola board
definitions, rename it (Gag!) "Force CPU Board" or some other meaningful
name, and alter the configuration of that definition to match the board.

>Please send your answers to : Ninane@fynu.ucl.ac.be.
>I will summarize.

     I have chosen NOT to email my reply but to post it to the group
instead, contrary to Alain's request.  I've done this for reasons of
expediency and publicity.  Frankly, I want my "role" as VMEexec-contact
to be known to readers of this newsgroup and I encourage you to email
questions and/or comments directly to me.  I will attempt to answer
each request in a timely, accurate, and complete fashion.  Please note,
however, that we frequently teach courses "on-site" at customer facilities
and, as such, I am out of town for an entire week.  (No, I don't have a
dial-up laptop I can take with me.  Maybe if this newsgroup keeps me
busy enough I'll be able to convince the powers that be of the need for
a dial-up machine. :-)
     Regardless, I will also try and respond to relevent questions within
this newsgroup.  In most cases I will simply email the information to the
requestor(s).

     Please feel free to contact me by email, phone, or Mail.

-- 
"Everybody who comes to a scientific lecture knows they are not going to
understand it, but maybe the lecturer has a nice, colored tie to look at."
Richard P. Feynman, "Q.E.D."
#include <disclaimer.h>
Ed Skinner, Technical Training, Motorola Inc., MicroComputer Division
2900 S Diablo Way, Tempe Az 85282, USA (602)438-3956
Internet: edski@phx.mcd.mot.com,  UUCPnet: noao!asuvax!mcdphx!edski