Emil J. Tywoniak
9abee44602
opt_expr: replace invert_map with signorm traversal
2026-05-22 18:41:49 +02:00
Emil J. Tywoniak
6fd7f5c02d
pmgen: hold sigmap pointer instead of owning it
2026-05-22 18:40:01 +02:00
Emil J. Tywoniak
5bfb631085
opt_expr: with -keepdc disable equality optimization rules that break when ports are sigmapped
2026-05-22 18:38:37 +02:00
Emil J. Tywoniak
aecc173f83
opt_dff: sigma harder, FfDataSigMapped
2026-05-22 18:38:37 +02:00
Emil J. Tywoniak
be7beaf91a
opt_dff: temporarily disable signorm due to muxtree traversal
2026-05-22 18:38:37 +02:00
Emil J. Tywoniak
68bb5c6b94
signorm: disable in passes that use swap_names
2026-05-22 18:37:58 +02:00
Emil J. Tywoniak
4d2a6f2b7a
opt_expr: fix invert_map
2026-05-22 18:37:58 +02:00
Emil J. Tywoniak
2f7d0913fc
opt_hier: disable signorm
2026-05-22 18:37:58 +02:00
Emil J. Tywoniak
bb2d6f0e2a
opt_merge_inc: re add initvals deletion
2026-05-22 18:37:13 +02:00
Emil J. Tywoniak
9d86a6636c
wreduce: fixup initvals after setPort
2026-05-22 18:37:13 +02:00
Jannis Harder
423c8be71b
WIP half broken snapshot
2026-05-22 18:37:11 +02:00
Emil J
e87a9bd9a7
Merge pull request #5888 from mikesinouye/pmux
...
opt_muxtree: reuse knowledge_t and pass by reference
2026-05-15 16:21:33 +00:00
Miodrag Milanović
36eceed720
Merge pull request #5862 from codexplorer-fish/cleaning-up-log-id
...
Cleaning up log_id()
2026-05-15 11:07:43 +00:00
Miodrag Milanovic
58df27ce7c
Refactor uses of log_id in pgm files
2026-05-14 12:21:32 +02:00
Mike Inouye
e20a57b09e
Reuse knowledge_t and pass by reference
2026-05-13 10:57:18 -07:00
Emil J. Tywoniak
3a150f2883
remove unused hashlib containers
2026-05-12 12:52:10 +02:00
Codexplorer
e41b969da2
Refactored uses of log_id()
2026-05-08 20:59:24 -07:00
Emil J. Tywoniak
7fa660fc60
share: remove -force
2026-05-04 21:34:19 +02:00
Emil J. Tywoniak
e0b833ac1a
opt_muxtree: dense knowledge
2026-04-24 11:07:58 +02:00
Emil J. Tywoniak
4abaca273e
opt_reduce: further optimization
2026-04-24 11:07:58 +02:00
Emil J. Tywoniak
4f4672d17b
muxpack: fix wide Y port handling
2026-03-19 00:12:49 +01:00
Emil J
c8f715fed8
Merge pull request #5664 from rocallahan/parallel-opt-clean
...
Parallelize `opt_clean` pass
2026-03-16 09:52:34 +00:00
Miodrag Milanovic
52533b0d1c
Update opt_lut_ins and stat for analogdevices and remove ecp5
2026-03-06 09:10:36 +01:00
Robert O'Callahan
9c51ba1b09
Reduce opt_clean parallelism
2026-03-06 02:20:16 +00:00
Robert O'Callahan
8d8c05b338
Fix OptCleanPass usage of CleanRunContext to avoid constructing extra KeepCache and ParallelDispatchThreadPool
2026-03-06 02:20:16 +00:00
Robert O'Callahan
32f5044eaf
Clarify "Not passing module as function argument" comment
...
This correct in terms of intent, it's just not fully enforced due to const laundering.
2026-03-06 02:20:16 +00:00
Emil J. Tywoniak
70cc2d67fd
opt_clean: refactor
2026-03-06 02:20:14 +00:00
Robert O'Callahan
3603cd52a0
Pass the module Subpool to rmunused_module_signals and parallelize that function
2026-03-06 02:20:08 +00:00
Robert O'Callahan
19a7c8fcf3
Pass the module Subpool to rmunused_module_cells and parallelize that function
2026-03-06 02:20:08 +00:00
Robert O'Callahan
8e044d1045
Pass the module Subpool to rmunused_module_init and parallelize that function
2026-03-06 02:20:06 +00:00
Robert O'Callahan
a7437c636d
Pass the toplevel thread pool to rmunused_module, create a Subpool, and parallelize remove_temporary_cells
2026-03-06 02:05:46 +00:00
Robert O'Callahan
887c32cb54
Create a toplevel ParallelDispatchThreadPool and parallelize keep_cache_t::scan_module() with it
2026-03-06 02:05:46 +00:00
Robert O'Callahan
72a21fe01d
Introduce RmStats struct to encapsulate removal statistics
...
Turns out this is not strictly necessary for this PR but it's
still a good thing to do and makes it clearer that the stats
are not modified in a possibly racy way.
2026-03-06 02:05:43 +00:00
Robert O'Callahan
c2bb7d6a82
Make keep_cache_t process all modules up-front instead of on-demand
...
We will want to query `keep_cache` from parallel threads. If we compute
the results on-demand, that means we need synchronization for cache
access in those queries, which adds complexity and overhead. Instead, prefill
the cache with the status of all relevant modules. Note that this doesn't
actually do more work --- we always consult `keep_cache` for all cells of
all selected modules, so scanning all those cells and determining the kept
status of all dependency modules is always required.
Later in this PR we're going to parallelize `scan_module` itself, and that's also
much easier to do when no other parallel threads are running.
2026-03-06 02:05:04 +00:00
Emil J. Tywoniak
07ec8708e4
share: use newcelltypes
2026-03-04 12:22:14 +01:00
Emil J. Tywoniak
9e59f05c25
newcelltypes: wrap design celltypes support
2026-03-04 12:22:14 +01:00
Emil J. Tywoniak
35ccaa60d7
newcelltypes: TurboCellTypes -> StaticCellTypes
2026-03-04 12:22:14 +01:00
Emil J. Tywoniak
6adc08b0e5
opt_expr: use newcelltypes
2026-03-04 12:22:14 +01:00
Emil J. Tywoniak
3671d577a0
opt_clean: use newcelltypes
2026-03-04 12:22:14 +01:00
Emil J
1717fa0180
Merge pull request #5663 from YosysHQ/emil/opt_expr-fix-pow-shift
...
opt_expr: fix const lhs of $pow to $shl
2026-02-05 13:09:01 +01:00
Emil J
8bbde80e02
Merge pull request #5631 from rocallahan/cleanup-compare-signals
...
Clean up `compare_signals()` in `opt_clean`
2026-02-04 17:45:05 +01:00
Emil J
992e64342c
Merge pull request #5621 from rocallahan/remove-opt-sort
...
Remove `Design::sort()` calls from optimization passes
2026-02-04 16:55:56 +01:00
Emil J. Tywoniak
3bfeaee8ca
opt_expr: fix const lhs of $pow to $shl
2026-02-03 11:59:00 +01:00
nella
8f6c4d40e4
Merge pull request #5623 from YosysHQ/nella/opt-dff-rewrite
...
opt_dff restructure.
2026-01-28 14:41:40 +01:00
nella
9367090763
OptDff more accurate ctrl/pattern desc.
2026-01-26 22:19:36 +01:00
nella
5803461c24
opt_dff pattern extraction.
2026-01-26 22:10:10 +01:00
nella
8576055dea
Fix tests.
2026-01-26 18:41:41 +01:00
nella
a75e0b2e92
opt_dff minor cleanup, added tests for comp var.
2026-01-26 14:24:01 +01:00
Robert O'Callahan
32e96605d4
Don't update used_signals for retained wires in rmunused_module_signals.
...
These updates should not be necessary. In fact, if they were necessary, this code
would be buggy, because the results would depend on the order in which wires are traversed:
If wire A is retained, which causes an update to `used_signals`, which then causes wire B
to be retained when it otherwise wouldn't be, then we would get different results depending
on whether A is visited before B.
These updates will also make it difficult to process these wires in parallel.
2026-01-24 03:41:18 +00:00
Robert O'Callahan
7d53d64a47
Make the call to compare_signals() easier to read.
...
The negation here is confusing. The intent of the code is "if `s1` is preferred
over `s2` as the canonical `SigBit` for this signal, make `s1` the canonical `SigBit`
in `assign_map`", so write the code that way instead of "if `s2` is not preferred
over `s1` ...".
This doesn't change any behavior now that `compare_signals()` is a total order,
i.e. `s1` is preferred over `s2`, `s2` is preferred over `s1`, or `s1` and `s2` are equal.
Now, when `s1` and `s2` are equal, we don't call `assign_map.add(s1)`, but that's
already a noop in that case.
2026-01-24 02:01:05 +00:00
