`dfflegalize` will emit NOT gates to drive control signals on flip-flops
when mapping to supported flip-flop polarities. Typically in a design
this will produce a number of NOT gates driven by the same signal. For
one reason or another ABC doesn't fully cancel this redundancy during
LUT mapping. Insert an explicit `opt_merge` pass to improve synthesis
QoR.
* Keep the previous behavior when no tcl script is used
* Do not treat "-" as a flag but as a positional argument
* Keep including <unistd.h> as it's also used for other functions (at
least for the emscripten build)
* Move the custom getopt implementation into the Yosys namespace to
avoid potential collisions
This expands the part-select tests with one additional module.
It specifically tests the different variants of the `peepopt`
optimizations `shiftadd` and `shiftmul`.
Not all these cases are actually transformed using `shiftadd`,
including them also checks if the correct variants are rejected.
- moved all selection and filtering logic to the match block
- applied less-verbose code suggestions
- removed constraint on number of bits in shift-amount
- added check for possible wrap-arround in the operation
Correct path to 011 source.
Also path for resources target.
Set timezone to 'Z' for faketime.
Not sure how to avoid needing to `make resources` before `make all` (or running
`make all` twice) in order to properly generate the presentation images.
Now includes usage output, (hopefully) generated by the tool during the docs build process so it will always be up to date.
Included in makefile as `docs/usage` target.
Also some updates/additions to the description text, esp `yosys-filterlib` and `yosys-smtbmc`.
Comment out the body text and instead include just the abstract and a download link.
Also orphan the pages so they aren't accessible except by direct link, or via search.
Add a separate shiftmul pattern to match on left shifts which implement
demuxing. This mirrors the right shift pattern matcher but is probably
best kept separate instead of merging the two into a single matcher.
In any case the diff of the two matchers should be easily readable.
The `opt_expr` pass running before `peepopt` can interfere with the
detection of a shiftmul pattern due to some of the bottom bits of the
shift amount being replaced with constant zero. Extend the detection to
cover those situations as well.