According to the XS1-L system specification it should be possible to "hot plug" links. I believe this also means it should be possible to dynamically allocate channels instead of statically declaring in the "main" of the program.
If that is the case is it possible to configure (for example) 2 startkits so that I can compile seperate "applications" and use xscope, xrun etc.?
From disassembling the 2-startkit example I think the process per node is:
1. Use GETR to get a channel resource
2. Use SETD to set the destination
3. Start issueing "HELLO" tokens until something responds.
From the documentation I read that it is also possible configure the switches but what I cannot find is how to indicate that I'd like to use, for instance, the "C" physical link to communicate.
So questions:
1. Is it possible to set up startkit XN files so that the <Node ID> entries are something different from "0".
2. If so, would that Node ID then correspond to the node id in the SETD destination?
3. Has anyone tried something like this before? If so, is there example code available?
Hotplugging links?
-
- XCore Addict
- Posts: 161
- Joined: Sun Jun 12, 2011 11:47 am
-
- XCore Addict
- Posts: 161
- Joined: Sun Jun 12, 2011 11:47 am
This wiki entry seems to be what I'm looking for!