| *** tpb has joined #vtr-dev | 00:00 | |
| *** digshadow has quit IRC | 02:15 | |
| *** digshadow has joined #vtr-dev | 02:42 | |
| *** digshadow has quit IRC | 03:11 | |
| *** digshadow has joined #vtr-dev | 03:16 | |
| *** digshadow has quit IRC | 03:20 | |
| *** digshadow has joined #vtr-dev | 03:20 | |
| elms | daveshah: question about vlog_to_model.py: If there is no interconnect in the pb_type (e.g. blackbox) does the model.xml need to include the models of subcomponents? It's not clear to me how port dependencies come into play such a case. | 17:55 |
|---|---|---|
| daveshah | elms: no, I don't think the models of subcomponents are needed if they will never be instantiated | 17:56 |
| elms | so should vlog_to_model.py generate an empty model.xml if blackbox is specified? | 17:57 |
| daveshah | elms: I think the correct output in that case would be a model.xml with only the blackbox model in it | 18:00 |
| daveshah | But I can't remember whether I ever implemented that correctly | 18:00 |
| elms | daveshah: I can look at it. Do you know of an example "only the blackbox model"? Or can you let me know what entries it would include. I'm still trying to grok the xml files fully | 18:03 |
| daveshah | elms: I would expect the model XML for a blackbox to look a bit like this | 18:05 |
| daveshah | https://www.irccloud.com/pastebin/qgLJRGah/ | 18:05 |
| tpb | Title: Snippet | IRCCloud (at www.irccloud.com) | 18:05 |
| elms | daveshah: ok thanks. I'll look at how to do that for the dsp48 | 18:08 |
| mithro | kem_: This was the bug I was talking about in the meeting -> https://github.com/verilog-to-routing/vtr-verilog-to-routing/issues/246 | 19:34 |
| tpb | Title: Invalid RR Graph constructed when pass_gate switches used · Issue #246 · verilog-to-routing/vtr-verilog-to-routing · GitHub (at github.com) | 19:34 |
| kem_ | mithro: Thanks | 19:36 |
| kem_ | mithro: For the mux modelling discussion see Section 2.3.2 of https://tspace.library.utoronto.ca/bitstream/1807/42733/3/Chiasson_Charles_RE_201311_MASc_thesis.pdf | 19:36 |
| kem_ | mithro: VPR models Figure 2.9 with the 'mux' switch type | 19:37 |
| mithro | kem_: Okay - I'll look into it | 19:38 |
| mithro | kem_: The way the ice40 seems to work is that you route onto "local tracks" from spans using pass gates, the local tracks can then be connected to logic pins | 19:39 |
| mithro | kem_: The local tracks can also be used to connect span wires into longer wires | 19:40 |
| mithro | kem_: I've been thinking of the local tracks as kind of like the "crossbar" that is normally at the front of the local tiles | 19:42 |
| *** digshadow has quit IRC | 21:28 | |
| *** digshadow has joined #vtr-dev | 21:32 | |
| *** digshadow has left #vtr-dev | 21:49 | |
| *** digshadow has joined #vtr-dev | 21:51 | |
| *** digshadow has quit IRC | 21:51 | |
| *** digshadow has joined #vtr-dev | 21:55 | |
| *** digshadow has quit IRC | 21:57 | |
| *** digshadow has joined #vtr-dev | 22:07 | |
| mithro | https://www.irccloud.com/pastebin/hHk1T98w/ | 22:40 |
| tpb | Title: Snippet | IRCCloud (at www.irccloud.com) | 22:40 |
| mithro | digshadow: ^ | 22:40 |
| mithro | digshadow: Looks like vpr is correct for complaining.... | 22:40 |
| digshadow | got it | 23:00 |
| *** benreynw1r has quit IRC | 23:44 | |
| *** benreynwar has joined #vtr-dev | 23:49 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!