Conny40 has quit [Remote host closed the connection]
rvalles_ has quit [Ping timeout: 265 seconds]
rvalles has joined #apicula
FabM has joined #apicula
<pepijndevos>
weeee pnr runs to completion
<pepijndevos>
I typod the IOB port directions
<pepijndevos>
Now I need to fix all the fun stuff such as command line arguments, constraint parsing, oh, and pass the real timing data...
<pepijndevos>
routing attosoc, the unrouted nets went down to 17 but now are just oscilating around 100 or so... :(
<pepijndevos>
yay, it finished!
<omnitechnomancer>
\o/
<omnitechnomancer>
how much utilisation is attosoc?
<pepijndevos>
with the new flow like 53%
<pepijndevos>
but yea routing seems to be struggling a lot. Not sure if it's because of lacking timing info for the moment, or because I should implement the mince stuff for synth_gowin
<daveshah>
it won't be the mince stuff - that helps placement not routing and anyway is usually only marginal
<pepijndevos>
ah ok
<daveshah>
it will be the poor timing data, delay estimates or other related stuff (like the ripup penalty)
<pepijndevos>
Yea I'm working on that now. I just hardcoded some imaginary numbers for the moment
<pepijndevos>
But yea placement in super fast so proper timing and estimates should improve it I hope
<omnitechnomancer>
fancy, using complex timing values
<pepijndevos>
welp bba doesn't even support floats
<pepijndevos>
Well, guess I'm fixing bba...
<daveshah>
is there any reason not to use an int32 picoseconds?
<pepijndevos>
not really, just that the vendor gives me floats
<pepijndevos>
but actually yea don't think it's worth the braincells to do it with floats