It doesn't matter hugely, but by default nextpnr synthesizes for 12MHz,
which doesn't force it to work too hard on the placement. By requesting
100MHz, it needs to try a bit harder on timing and gives results that
are a bit closer to the fully constrained outcomes.
To help divvy up the sources better, the build now makes a libpath
consisting of all directories that have bsv files in them, with a few
exceptions: hardware subdirs are target-specific so only get used if
they're the current build target. Experiments are random crap so get
the same treatment. And the 'sim' dir is only test helpers, so they
only get pulled in by tests.
Yosys doesn't understand Verilog-2001 port aliases. Unfortunately
bsc uses those to represent inout ports because it's the only way
to represent a particular kind of shared bus in Verilog source code.
Thankfully, a kind soul at Bluespec Inc made a perl script that
transforms the port alias construct into regular verilog-1995, which
works fine in cases like mine where the only user of the inout port
is a TriState module which tears it apart into separate
input/output/enable signals for the rest of bsc to work with.
A delay line takes a write and echoes it back N cycles later,
with N fixed at compile time. It's a handy primitive to have
when wrapping Verilog blackbox modules because the blackbox
often specifies something like having 2 cycles of latency,
and so you need to bubble the fact that a write occurred 2
cycles ago through to the output so that you can wire up the
right implicit conditions.
Partial synth is handy when writing gnarly Bluespec modules, because it
lets you inspect the Verilog output of the Bluespec compiler as well as
Yosys's compile output at various stages of synthesis, to see if things
are being produced the way you expect.
David Anderson