[Openib-windows] RE: IBDM - IB DataModel build issues fixed

[Eitan Zahavi]

Hi Hal,

> Why does the TOPSYSTEM need to be identified ? Is it just a way of not
> having it as the first file in the system definition ?
> 
> The IBNL document mentions IBADM. What is IBADM ?
[EZ] I'm sorry about that. It is really IBDM. I attach here the updated
document. Also removed any IBADM remedies from the install path.

> 
> It also mentions IB netlist and topology file. What is the IB netlist ?
[EZ] IBNL is IB NetList - it is the internal system connectivity file. The
term "net list" is very common in the CAD world for describing connectivity
information.

> Where does the topology file come from ? 
[EZ] The user is supposed to provide that. Normally when large clusters are
built the same information is used for labeling the cables.

Is there a topology file which
> uses system names (and requires underlying system definition files (to
> be predefined)) ?
[EZ] I am not sure what you mean by "Is there"? 
Yes the topology file will use systems that are defined by IBNL and list
their inter connections. I already provided an example:

MT2400 mySwitch
   P1 -> Cougar H-1 P2
   P2 -> Cougar H-2 P1 
   
Is a simple topology file describing two Cougars (but you might use other
HCA names of course) to a 24 port switch system named MT2400.

> 
> Also, can you list the (current and anticipated) consumers of this data
> model ? 
[EZ] Good question:
ibnlparse - a simple parser and checker for IBNL files only. 

I presume the IBMgtSim is one. I also see: ibnlparse, ibdmchk,
> ibdmtr, and topotest. Are there others ? Is there any description of
> these ?
> 
> Thanks.
> 
> -- Hal

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/ofw/attachments/20050705/f418f1b9/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: IBNL Specification.doc
Type: application/msword
Size: 96768 bytes
Desc: not available
URL: <http://lists.openfabrics.org/pipermail/ofw/attachments/20050705/f418f1b9/attachment.doc>