d9b5bf65e8
glucose2: Don't try to justify toplevel assignments
...
Doing so isn't necessary and would occasionally result in violated
assertions within glucose2.
2024-09-07 13:16:20 +02:00
d1aa15b0c8
glucose2: Create initial extra jlevel entries
...
My understanding is that jlevel needs to have this extra space as there
is existing code that assumes it is safe to index into it at index
decisionLevel()+1.
2024-08-19 14:15:49 +02:00
1e02c4327c
glucose2: Allow user defined marking of limited depth cones
...
This does require using the version of uncheckedEnqueue that does check
whether the target lit is active in the current round.
2024-08-19 14:06:17 +02:00
f36e35f735
glucose2: Improvements and fixes for proof tracing
...
This fills some gaps and fixes a few edge cases with the new proof
tracing callback API.
2024-08-19 13:47:53 +02:00
e2749d5c90
glucose2: Allow re-definition of gates
...
This enables e.g. rewriting the circuit based on discovered
equivalences.
This avoids re-introducing doubly linked lists for the fanout by instead
keeping a vector of variables whose fanout lists contain gate inputs
that were changed to a different literal. The first time propagation
runs with that vector nonempty, it deduplicates that vector and then
traverses the affected lists and moves every changed gate input into the
correct list.
2024-08-19 13:45:19 +02:00
b7ea0adc9f
glucose2: Add option to include inner gates in the justification model
2024-08-19 13:38:14 +02:00
a6512064bc
First version of glucose2 proof tracing API
2024-08-05 12:02:16 +02:00
28d955ca97
Merge pull request #35 from lf-/jade/fix-gitarchive
...
Fix archive reproducibility
2024-07-08 19:02:21 +02:00
04f50406fd
Fix archive reproducibility
...
The git archive export-subst option does not have consistent results
over time, since the abbreviated commit hash can get longer over time.
Instead, let's export the full commit hash.
This was found in an audit of source archive reproducibility in nixpkgs:
```
~ » expected=$(nix-store -r /nix/store/4j9rj4m6akjskp0f7k923qff817k6hv5-source)
actual=$(nix-prefetch-url --print-path --unpack --name source 896e5e7dedhttps://cache.nixos.org '...
warning: you did not specify '--add-root'; the result might be removed by the garbage collector
--- /nix/store/4j9rj4m6akjskp0f7k923qff817k6hv5-source
+++ /nix/store/v7ms5ghibzi8pk71nzlhvsbn7a0rdpy7-source
│ --- /nix/store/4j9rj4m6akjskp0f7k923qff817k6hv5-source/.gitcommit
├── +++ /nix/store/v7ms5ghibzi8pk71nzlhvsbn7a0rdpy7-source/.gitcommit
│ @@ -1 +1 @@
│ -896e5e7de
│ +896e5e7ded
│ ├── stat {}
│ │ @@ -1,7 +1,7 @@
│ │
│ │ - Size: 10 Blocks: 1 IO Block: 512 regular file
│ │ + Size: 11 Blocks: 1 IO Block: 512 regular file
│ │ Device: 0,24 Access: (0444/-r--r--r--) Uid: ( 0/ root) Gid: ( 0/ root)
│ │
│ │ Modify: 1970-01-01 00:00:01.000000000 +0000
```
2024-06-27 13:41:10 -07:00
237d81397f
Modify include guards in cmd.c so that Windows compilers don't compile Unix-only code.
2024-05-08 08:26:54 +02:00
03da96f12f
Patch for lack of `system` on WASM
2024-04-23 13:49:12 +02:00
b502f00222
Fix prototype mismatch for `Gia_ManSimRsb`
2024-04-22 17:39:32 +02:00
208b48667f
Merge branch 'master' of https://github.com/berkeley-abc/abc into yosys-experimental
2024-04-22 16:33:29 +02:00
2c52e3f969
Support out of tree builds
2024-04-18 04:49:09 +01:00
8a174ee865
Merge pull request #287 from gadfort/correct-prefix
...
ensure initial library writing also honors prefix
2024-04-16 06:22:46 -07:00
de060a26ad
ensure initial library writing also honors prefix
2024-04-16 08:58:28 -04:00
bc725b85de
Bug fix in CNF generation for &glucose (three more places).
2024-04-15 20:29:38 -07:00
2d6b5c9adc
Bug fix in CNF generation for &glucose.
2024-04-15 20:25:43 -07:00
99e0e37da6
Added switch -p in "read_lib" to skip writing cell prefix.
2024-04-14 09:51:00 -07:00
682480e5e9
Merge pull request #285 from gadfort/add-lib-merging
...
add library merging flag to read_lib
2024-04-14 09:39:14 -07:00
935c6a875d
add missing flag to read_lib help
2024-04-12 13:49:44 -04:00
1d90cafd54
add library merging flag to read_lib
...
Signed-off-by: Peter Gadfort <gadfort@zeroasic.com>
2024-04-12 13:27:58 -04:00
b10d000c7a
Merge pull request #284 from aletempiac/acd66
...
LUT structure mapping
2024-04-11 12:05:12 -07:00
045803dcb8
Merge remote-tracking branch 'origin/master' into acd66
2024-04-11 19:02:29 +02:00
0c905f873b
Fixes
2024-04-11 19:01:05 +02:00
6052d10fde
Adding new command if -U for 2-LUT decompositions under delay profile
2024-04-11 15:45:37 +02:00
e8924e5534
Fixes and improvements
2024-04-11 15:44:52 +02:00
5b49724fcc
removing acd666
2024-04-11 15:43:22 +02:00
1446b7549f
Merge pull request #29 from thorpej/dev/pkgsrc-patch-NetBSD-1
...
Fix building on NetBSD
2024-04-11 14:48:28 +02:00
078afe9faa
Merge pull request #31 from davidar/fix-55
...
Fix Assertion using &if: `pCutSet->nCuts > 0'
2024-04-11 14:41:43 +02:00
ccc02c4400
Merge pull request #30 from povik/&mfs-fixes
...
Address some `&mfs` crashes
2024-04-11 14:41:24 +02:00
ca78f5e6e5
Bug fix in the resub engine.
2024-04-11 05:05:52 -07:00
32bc1d4ab2
Cleaning and generalizing code
2024-04-11 11:31:28 +02:00
64fea5c4c2
Improving the performance and quality of acd66
2024-04-10 18:43:52 +02:00
6b5ebb3e76
Removing assertion when decomposing into LUTs smaller than 6
2024-04-10 18:42:52 +02:00
8f3447800c
Support again decompositions into luts smaller than 6
2024-04-02 11:25:03 +02:00
accf50468a
Apply patch to If_ObjPerformMappingChoice too
2024-04-01 10:03:10 +10:00
316eec6d3f
Fix Assertion using &if: `pCutSet->nCuts > 0'
2024-04-01 09:40:41 +10:00
6e1653426f
Switch to randomly select one choice.
2024-03-28 16:22:06 +08:00
a2cb5eb4e3
Adding command &pms to print miter status.
2024-03-25 23:39:03 +08:00
1f72ffce79
Improving ACD performance with bail-out conditions
2024-03-25 14:23:43 +01:00
b0d2ff1c63
Exact synthesis using NAND-gates.
2024-03-24 00:10:08 +09:00
1107634fa6
&mfs: Make it no biggie when a network is all blackboxes, no whiteboxes
2024-03-22 22:47:40 +01:00
d7fc8fe98f
&mfs: Handle blackboxes robustly
...
When the network is being handed over to the "sfm" core, all blackboxes
are modeled by inserting new PIs, POs, and those being connected by
buffers to the nodes representing the CIs, COs. Make two changes:
* Robustly deny the fake PIs from being considered when shopping for
LUT fanin substitutions. Such reconnection occurring would trip up
the code reintegrating the result.
* Make sure the buffer connecting the fake-PO to the CO doesn't get
rewritten as part of the `mfs` transformation, and extend this
protection to any whitebox models.
2024-03-22 22:43:08 +01:00
fda490235e
&mfs: Fix issues with traversal when re-importing network
...
The former implementation of `Sfm_NtkDfs` was trying to serialize the
network while ordering all box inputs ahead of the box outputs. This is
sometimes impossible, leaving the result unordered, which led to crashes
in the `&mfs` code when it was reintegrating the result into the GIA
structure:
ABC: Assertion failed: iLitNew >= 0 (src/aig/gia/giaMfs.c: Gia_ManInsertMfs: 388)
With a small change to `Gia_ManInsertMfs` which does the reintegration
we don't really need the ordering to see through boxes, ordering on the
paths between boxes is sufficient. Relaxing the ordering requirement, we
make `Sfm_NtkDfs` robust.
2024-03-22 22:01:45 +01:00
b83985c25b
Add `&mfs -r` for re-import testing
2024-03-22 21:42:12 +01:00
6aacf524aa
Performance improvement and fixes
2024-03-22 19:19:35 +01:00
8a314db8dc
Bug fix
2024-03-22 15:39:52 +01:00
783a5404a2
Fixing Windows makefile.
2024-03-19 09:57:52 +09:00
2c0943ff62
Fixiing compiler problem on Windows.
2024-03-19 09:34:20 +09:00
Jannis Harder
Emil J
Jade Lovelace
William D. Jones
Martin Povišer
Roland Coeurjoly
alanminko
Peter Gadfort
Alan Mishchenko
aletempiac
Miodrag Milanović
N. Engelhardt
David A Roberts