AutonamePass rescanned every selected cell/wire from scratch on every
iteration, and naming only propagates one hop per iteration, so cost
scaled with O(iterations x module_size). On a large, fully flattened
netlist with long dependency chains this meant multi-hour runtime and
40+ GB RSS. Likely the same issue as #5394, #4509 and #2816.
Replace the full-rescan loop with a ModuleAutonamer that keeps a
persistent worklist and only recomputes an object's proposal when a
direct neighbor was just renamed, using precomputed adjacency. Same
round-by-round batching as before, so the final naming is unchanged.
Adds tests/various/autoname_scaling.sh, which checks a 10000-cell
propagation chain still autonames within a time budget the old
algorithm could not meet.
Rewrite power-of-two indexed word selects to $bmux when the shift
amount already carries the scale as low zero bits.
Keep the rule to non-overlapping selections and bound the generated
mux ways. Add regressions for aligned shifts, padding, signed
extension, and shiftmul handoff cases.
This uses the environment variable `YOSYS_PLUGIN_PATH` to provide multiple colon-delimited search paths for native plugins in a similar manner to `PATH` for executables and `PYTHONPATH` for Python modules.
This addresses https://github.com/YosysHQ/yosys/issues/2545, allowing Yosys to be better packaged in non-FHS environments such as Nix.
This fixes some edge cases the previous version didn't handle properly
by simplifying the logic of determining directly driven wires and
representatives to use as buffer inputs.
Was previously the number of proposed renames, but since renames can be skipped this causes the final count to differ from the number of actually renamed objects.
Check counts in `tests/various/autoname.ys`.
This is a complete rewrite of the RTLIL-kernel-side bufnorm code. This
is done to support inout ports and undirected connections as well as to
allow removal of cells while in bufnorm mode.
This doesn't yet update the (experimental) `bufnorm` pass, so to
manually test the new kernel functionality, it is important to only use
`bufnorm -update` and `bufnorm -reset` which rely entirely on the kernel
functionality. Other modes of the `bufnorm` pass may still fail in the
presence of inout ports or undirected connections.
* When used with -tempinduct mode, -seq <N> causes assertions to be
ignored in the first N steps. While this has uses for reset modelling,
for these test cases it is unnecessary and could lead to failures
slipping through uncaught
